records.yaml

Important

From ATS 10 we have moved from the old records.config format style in flavour of the new YAML style. Please check Converting records.config to records.yaml for information about migrating from the old style to the new YAML base config.

The records.yaml file (by default, located in /usr/local/etc/trafficserver/) is a YAML base configuration file used by the Traffic Server software. After you modify records.yaml, run the command traffic_ctl config reload to apply the changes.

Note

The configuration directory, containing the SYSCONFDIR value specified at build time relative to the installation prefix, contains Traffic Server configuration files. The $TS_ROOT environment variable can be used alter the installation prefix at run time. The directory must allow read/write access for configuration reloads.

YAML structure

All fields are located inside the records` root node. ATS supports reading multiple documents from the same YAML stream, subsequent documents overrides earlier fields. If there is any error reading a particular node the default value(s) will be used.

records:
  diags:
   debug:
      enabled: 0
      tags: http|dns
 # ...
 # rest of the fields.
 # ...

Important

Internally, ATS uses record names as Configuration Variables.

Data Type

There is no need to manually set the record type, it can be done if desired. The types accepted are:

Type	Description
float	Floating point, expressed as a decimal number without units or exponents.
int	Integers, expressed with or without unit prefixes (as described below).
str	String of characters up to the first newline. No quoting necessary.

Non core records

Records that aren’t part of the core ATS needs to set the field type, this for now is the only way to know the field type.

We expect non core records to set the type (!!int, !!float, etc).

records:
   plugin_x:
      my_field_1: !!int '1'
      my_field_2: !!float '1.2'
      my_field_3: 'my string'

Values

The field_value must conform to the variable’s type. For str, this is simply any character data until the first newline.

For integer (int) fields, values are expressed as any normal integer, e.g. 32768. They can also be expressed using more human readable values using standard unit prefixes, e.g. 32K. The following prefixes are supported for all int type configurations:

Prefix	Description	Equivalent in Bytes
K	Kilobytes	1,024 bytes
M	Megabytes	1,048,576 bytes (10242)
G	Gigabytes	1,073,741,824 bytes (10243)
T	Terabytes	1,099,511,627,776 bytes (10244)

Floating point variables (float) must be expressed as a regular decimal number. Unit prefixes are not supported, nor are alternate notations (scientific, exponent, etc.).

Additional Attributes

Deprecated

A variable marked as Deprecated is still functional but should be avoided as it may be removed in a future release without warning.

Reloadable

A variable marked as Reloadable can be updated via the command:

traffic_ctl config reload

This updates configuration parameters without restarting Traffic Server or interrupting the processing of requests.

Overridable

A variable marked as Overridable can be changed on a per-remap basis using plugins (like the Configuration Remap Plugin), affecting operations within the current transaction only. Remap config files still uses the legacy records.config style.

Examples

In the following example, the field proxy_name is a str datatype with the value my_server. This means that the name of the Traffic Server proxy is my_server.

records:
   proxy_name: my_server

If the server name should be that_server the line would be:

records:
   proxy_name: that_server

In the following example, the field is a yes/no flag. A value of 0 (zero) disables the option; a value of 1 enables the option.

records:
   arm:
      enabled: 0

In the following example, the field sets the time to wait for a DNS response to 10 seconds.

records:
   hostdb:
      lookup_timeout: 10

In the following example the field sets the field with a float value.

records:
   exec_thread:
      autoconfig:
         scale: 1.0

The last examples configures a 64GB RAM cache, using a human readable prefix.

records:
   cache:
      ram_cache:
         size: 64G

Environment Overrides

Every records.yaml configuration variable can be overridden by a corresponding environment variable. This can be useful in situations where you need a static records.yaml but still want to tweak one or two settings. The override variable is formed by converting the records.yaml variable name to upper case, and replacing any dot separators with an underscore.

Overriding a variable from the environment is permanent and will not be affected by future configuration changes made in records.yaml or applied with traffic_ctl.

For example, we could override the proxy.config.product_company variable like this:

$ PROXY_CONFIG_PRODUCT_COMPANY=example traffic_server &
$ traffic_ctl config get proxy.config.product_company

Configuration Variables

The following list describes the configuration variables available in the records.yaml file.

System Variables

proxy.config.product_company

Scope:

CONFIG

Type:

STRING

Default:

Apache Software Foundation

yaml:

records:
  product_company: Apache Software Foundation

The name of the organization developing Traffic Server.

proxy.config.product_vendor

Scope:

CONFIG

Type:

STRING

Default:

Apache

yaml:

records:
  product_vendor: Apache

The name of the vendor providing Traffic Server.

proxy.config.product_name

Scope:

CONFIG

Type:

STRING

Default:

|TS|

yaml:

records:
  product_name: '|TS|'

The name of the product.

proxy.config.proxy_name

Scope:

CONFIG

Type:

STRING

Default:

build_machine

Reloadable:

Yes

yaml:

records:
  proxy_name: build_machine

The name of the Traffic Server node.

proxy.config.bin_path

Scope:

CONFIG

Type:

STRING

Default:

bin

yaml:

records:
  bin_path: bin

The location of the Traffic Server bin directory.

proxy.config.memory.max_usage

Scope:

CONFIG

Type:

INT

Default:

0

Units:

bytes

yaml:

records:
  memory:
    max_usage: 0

Throttle incoming connections if resident memory usage exceeds this value. Setting the option to 0 disables the feature.

proxy.config.syslog_facility

Scope:

CONFIG

Type:

STRING

Default:

LOG_DAEMON

yaml:

records:
  syslog_facility: LOG_DAEMON

The facility used to record system log files. Refer to Understanding Traffic Server Logs for more in-depth discussion of the contents and interpretations of log files.

proxy.config.output.logfile.name

Scope:

CONFIG

Type:

STRING

Default:

traffic.out

yaml:

records:
  output:
    logfile:
      name: traffic.out

This is used for log rolling configuration so Traffic Server knows the path of the output file that should be rolled. This configuration takes the name of the file receiving traffic_server process output that is set via the --bind_stdout and --bind_stderr command-line options. proxy.config.output.logfile.name is used only to identify the name of the output file for log rolling purposes and does not override the values set via --bind_stdout and --bind_stderr.

If a filename is passed to this option, then it will be interpreted relative to proxy.config.log.logfile_dir. If a different location is desired, then pass an absolute path to this configuration.

proxy.config.output.logfile_perm

Scope:

CONFIG

Type:

STRING

Default:

rw-r–r–

yaml:

records:
  output:
    logfile_perm: rw-r--r--

The log file permissions for the file receiving Traffic Server output, the path of which is configured via the --bind_stdout and --bind_stderr command-line options. The standard UNIX file permissions are used (owner, group, other). Permissible values are:

Value	Description
-	No permissions.
r	Read permission.
w	Write permission.
x	Execute permission.

Permissions are subject to the umask settings for the Traffic Server process. This means that a umask setting of 002 will not allow write permission for others, even if specified in the configuration file. Permissions for existing log files are not changed when the configuration is modified.

proxy.config.output.logfile.rolling_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  output:
    logfile:
      rolling_enabled: 0

Specifies how the output log is rolled. You can specify the following values:

Value	Description
0	Disables output log rolling.
1	Enables output log rolling at specific intervals (specified with the proxy.config.output.logfile.rolling_interval_sec variable). The clock starts ticking on Traffic Server boot.
2	Enables output log rolling when the output log reaches a specific size (specified with proxy.config.output.logfile.rolling_size_mb).
3	Enables output log rolling at specific intervals or when the output log reaches a specific size (whichever occurs first).

proxy.config.output.logfile.rolling_interval_sec

Scope:

CONFIG

Type:

INT

Default:

3600

Units:

seconds

Reloadable:

Yes

yaml:

records:
  output:
    logfile:
      rolling_interval_sec: 3600

Specifies how often the output log is rolled, in seconds. The timer starts on Traffic Server startup.

proxy.config.output.logfile.rolling_size_mb

Scope:

CONFIG

Type:

INT

Default:

100

Units:

megabytes

Reloadable:

Yes

yaml:

records:
  output:
    logfile:
      rolling_size_mb: 100

Specifies at what size to roll the output log at.

proxy.config.output.logfile.rolling_min_count

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  output:
    logfile:
      rolling_min_count: 0

Specifies the minimum count of rolled output logs to keep. This value will be used to decide the order of auto-deletion (if enabled). A default value of 0 means auto-deletion will try to keep output logs as much as possible. See Log Rotation and Retention for guidance.

Thread Variables

proxy.config.exec_thread.autoconfig.enabled

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  exec_thread:
    autoconfig:
      enabled: 1

When enabled (the default, 1), Traffic Server scales threads according to the available CPU cores. See the config option below.

proxy.config.exec_thread.autoconfig.scale

Scope:

CONFIG

Type:

FLOAT

Default:

1.0

yaml:

records:
  exec_thread:
    autoconfig:
      scale: 1.0

Factor by which Traffic Server scales the number of threads. The multiplier is usually the number of available CPU cores. By default this is scaling factor is 1.0.

proxy.config.exec_thread.limit

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  exec_thread:
    limit: 2

The number of threads Traffic Server will create if proxy.config.exec_thread.autoconfig.enabled is set to 0, otherwise this option is ignored.

proxy.config.exec_thread.listen

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  exec_thread:
    listen: 0

If enabled (1) all the exec_threads listen for incoming connections. proxy.config.accept_threads should be disabled to enable this variable.

proxy.config.accept_threads

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  accept_threads: 1

The number of accept threads. If disabled (0), then accepts will be done in each of the worker threads.

accept_threads	exec_thread.listen	Effect
0	0	All worker threads accept new connections and share listen fd.
1	0	New connections are accepted on a dedicated accept thread and distributed to worker threads in round robin fashion.
0	1	All worker threads listen on the same port using SO_REUSEPORT. Each thread has its own listen fd and new connections are accepted on all the threads.

By default, proxy.config.accept_threads is set to 1 and proxy.config.exec_thread.listen is set to 0.

proxy.config.thread.default.stacksize

Scope:

CONFIG

Type:

INT

Default:

1048576

yaml:

records:
  thread:
    default:
      stacksize: 1048576

Default thread stack size, in bytes, for all threads (default is 1 MB).

proxy.config.exec_thread.affinity

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  exec_thread:
    affinity: 1

Bind threads to specific processing units.

Value	Effect
0	Assign threads to machine.
1	Assign threads to NUMA nodes [default].
2	Assign threads to sockets.
3	Assign threads to cores.
4	Assign threads to processing units.

Note

This option only has an affect when Traffic Server has been compiled with --enable-hwloc.

proxy.config.system.file_max_pct

Scope:

CONFIG

Type:

FLOAT

Default:

0.9

yaml:

records:
  system:
    file_max_pct: 0.9

Set the maximum number of file handles for the traffic_server process as a percentage of the fs.file-max proc value in Linux. The default is 90%.

proxy.config.crash_log_helper

Scope:

CONFIG

Type:

STRING

Default:

traffic_crashlog

yaml:

records:
  crash_log_helper: traffic_crashlog

This option directs traffic_server to spawn a crash log helper at startup. The value should be the path to an executable program. If the path is not absolute, it is located relative to configured bin directory. Any user-provided program specified here must behave in a fashion compatible with traffic_crashlog. Specifically, it must implement the traffic_crashlog --wait behavior.

This setting not reloadable because the helper must be spawned before traffic_server drops privilege. If this variable is set to NULL, no helper will be spawned.

proxy.config.restart.stop_listening

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  restart:
    stop_listening: 0

This option specifies whether Traffic Server should close listening sockets while shutting down gracefully.

Value	Description
0	Listening sockets will be kept open.
1	Listening sockets will be closed when Traffic Server starts shutting down.

proxy.config.stop.shutdown_timeout

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  stop:
    shutdown_timeout: 0

The shutdown timeout(in seconds) to apply when stopping Traffic Server, in which ATS can initiate graceful shutdowns. In order to effect graceful shutdown, the value specified should be greater than 0. Value of 0 will not effect an abrupt shutdown. Abrupt shutdowns can be achieved with out specifying –drain; (traffic_ctl server stop /restart). Stopping Traffic Server here means sending traffic_server a signal either by bin/trafficserver stop or kill.

proxy.config.thread.max_heartbeat_mseconds

Scope:

CONFIG

Type:

INT

Default:

60

Units:

milliseconds

yaml:

records:
  thread:
    max_heartbeat_mseconds: 60

Set the maximum heartbeat in milliseconds for threads, ranges from 0 to 1000.

This controls the maximum amount of time the event loop will wait for I/O activity. On a system that is not busy, this option can be set to a higher value to decrease the spin around overhead. If experiencing unexpected delays, setting a lower value should improve the situation. Note that this setting should only be used by expert system tuners, and will not be beneficial with random fiddling.

Network

proxy.config.net.additional_accepts

Scope:

CONFIG

Type:

INT

Default:

-1

Reloadable:

Yes

yaml:

records:
  net:
    additional_accepts: -1

This config addresses an issue that can sometimes happen if threads are caught in a net accept while loop, become busy exclusviely accepting connections, and are prevented from doing other work. This can cause an increase in latency and average event loop time. When set to 0, a thread accepts only 1 connection per event loop. When set to any other positive integer x, a thread will accept up to x+1 connections per event loop. When set to -1 (default), a thread will accept connections as long as there are connections waiting in its listening queue.is equivalent to “accept all”, and setting to 0 is equivalent to “accept one”.

proxy.config.net.connections_throttle

Scope:

CONFIG

Type:

INT

Default:

30000

yaml:

records:
  net:
    connections_throttle: 30000

The total number of client and origin server connections that the server can handle simultaneously. This is in fact the max number of file descriptors that the traffic_server process can have open at any given time. Roughly 10% of these connections are reserved for origin server connections, i.e. from the default, only ~27,000 client connections can be handled. This should be tuned according to your memory size, and expected work load. If this is set to 0, the throttling logic is disabled.

proxy.config.net.max_connections_in

Scope:

CONFIG

Type:

INT

Default:

30000

yaml:

records:
  net:
    max_connections_in: 30000

The total number of client requests that Traffic Server can handle simultaneously. This should be tuned according to your memory size, and expected work load (network, cpu etc). This limit includes both idle (keep alive) connections and active requests that Traffic Server can handle at any given instant. The delta between proxy.config.net.max_connections_in and proxy.config.net.max_requests_in is the amount of maximum idle (keepalive) connections Traffic Server will maintain. If this is set to 0, the throttling logic is disabled.

proxy.config.net.per_client.max_connections_in

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  net:
    per_client:
      max_connections_in: 0

The total number of concurrent client connections that Traffic Server will accept from a given client IP address. Any received connections from a client beyond this limit will be immediately closed. Once the number of concurrent client connections drops below this configured value, Traffic Server will begin accepting new connections from that IP while the number of concurrent connections remains below this limit. A value of 0 disables the per client concurrent connection limit.

proxy.config.http.per_client.connection.alert_delay

Scope:

CONFIG

Type:

INT

Default:

60

Units:

seconds

Reloadable:

Yes

yaml:

records:
  http:
    per_client:
      connection:
        alert_delay: 60

Throttle alerts per client IP address to be no more often than this many seconds. Summary data is provided per alert to allow log scrubbing to generate accurate data.

proxy.config.net.max_requests_in

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  net:
    max_requests_in: 0

The total number of concurrent requests or active client connections that the Traffic Server can handle simultaneously. This should be tuned according to your memory size, and expected work load (network, cpu etc). When set to 0, active request tracking is disabled and max requests has no separate limit and the total connections follow proxy.config.net.connections_throttle

proxy.config.net.default_inactivity_timeout

Scope:

CONFIG

Type:

INT

Default:

86400

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  net:
    default_inactivity_timeout: 86400

The connection inactivity timeout (in seconds) to apply when Traffic Server detects that no inactivity timeout has been applied by the HTTP state machine. When this timeout is applied, the proxy.process.net.default_inactivity_timeout_applied metric is incremented.

Note that this configuration is overridable. While most overridable configurations conceptually apply to specific transactions, default_inactivity_timeout is a connection level concept. This is not necessarily a problem, but it does mean that care must be taken when applying the override to consider that all transactions in the connection which has this timeout overridden will be impacted by the override. For instance, if the default inactivity timeout is being overridden via a Configuration Remap Plugin rule in remap.config, then all transactions for that connection will be impacted by the override, not just the ones matching that remap.config rule.

See Timeout Settings for more discussion on Traffic Server timeouts.

Note that for QUIC this will be ignored and proxy.config.quic.no_activity_timeout_in should be used instead.

proxy.config.net.inactivity_check_frequency

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  net:
    inactivity_check_frequency: 1

How frequent (in seconds) to check for inactive connections. If you deal with a lot of concurrent connections, increasing this setting can reduce pressure on the system.

proxy.config.incoming_ip_to_bind

Scope:

CONFIG

Type:

STRING

Default:

0.0.0.0 [::]

yaml:

records:
  incoming_ip_to_bind: 0.0.0.0 [::]

Controls the global default IP addresses to which to bind proxy server ports. The value is a space separated list of IP addresses, one per supported IP address family (currently IPv4 and IPv6).

Unless explicitly specified in proxy.config.http.server_ports, the server port will be bound to one of these addresses, selected by IP address family. The built in default is any address. This is used if no address for a family is specified. This setting is useful if most or all server ports should be bound to the same address.

Note

This is ignored for inbound transparent server ports because they must be able to accept connections on arbitrary IP addresses.

Example

Set the global default for IPv4 to 192.168.101.18 and leave the global default for IPv6 as any address

records:
  incoming_ip_to_bind: 192.168.101.18

Example

Set the global default for IPv4 to 191.68.101.18 and the global default for IPv6 to fc07:192:168:101::17

records:
  incoming_ip_to_bind: "192.168.101.18 [fc07:192:168:101::17]"

proxy.config.outgoing_ip_to_bind

Scope:

CONFIG

Type:

STRING

Default:

0.0.0.0 [::]

yaml:

records:
  outgoing_ip_to_bind: 0.0.0.0 [::]

This controls the global default for the local IP address for outbound connections to origin servers. The value is a list of space separated IP addresses, one per supported IP address family (currently IPv4 and IPv6).

Unless explicitly specified in proxy.config.http.server_ports, one of these addresses, selected by IP address family, will be used as the local address for outbound connections. This setting is useful if most or all of the server ports should use the same outbound IP addresses.

Note

This is ignored for outbound transparent ports as the local outbound address will be the same as the client local address.

Example

Set the default local outbound IP address for IPv4 connections to 192.168.101.18.

records:
  outgoing_ip_to_bind: 192.168.101.18

Example

Set the default local outbound IP address to 192.168.101.17 for IPv4 and fc07:192:168:101::17 for IPv6.

records:
  outgoing_ip_to_bind: "192.168.101.17 [fc07:192:168:101::17]"

proxy.config.net.event_period

Scope:

CONFIG

Type:

INT

Default:

10

yaml:

records:
  net:
    event_period: 10

How often, in milli-seconds, to schedule IO event processing. This is unlikely to be necessary to tune, and we discourage setting it to a value smaller than 10ms (on Linux).

proxy.config.net.accept_period

Scope:

CONFIG

Type:

INT

Default:

10

yaml:

records:
  net:
    accept_period: 10

How often, in milli-seconds, to schedule accept() processing. This is unlikely to be necessary to tune, and we discourage setting it to a value smaller than 10ms (on Linux).

proxy.config.net.retry_delay

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

yaml:

records:
  net:
    retry_delay: 10

How long to wait until we retry various events that would otherwise block the network processing threads (e.g. locks). We discourage setting this to a value smaller than 10ms (on Linux).

proxy.config.net.throttle_delay

Scope:

CONFIG

Type:

INT

Default:

50

Reloadable:

Yes

yaml:

records:
  net:
    throttle_delay: 50

When we trigger a throttling scenario, this how long our accept() are delayed.

Management

proxy.config.admin.user_id

Scope:

CONFIG

Type:

STRING

Default:

nobody

yaml:

records:
  admin:
    user_id: nobody

Designates the non-privileged account to run the traffic_server process as, which also has the effect of setting ownership of configuration and log files.

If the user_id is prefixed with pound character (#), the remainder of the string is considered to be a numeric user identifier. If the value is set to #-1, Traffic Server will not change the user during startup.

Important

Attempting to set this option to root or #0 is now forbidden, as a measure to increase security. Doing so will cause a fatal failure upon startup in traffic_server. However, there are two ways to bypass this restriction:

• Specify -DBIG_SECURITY_HOLE in CXXFLAGS during compilation.

• Set the user_id=#-1 and start trafficserver as root.

proxy.config.admin.api.restricted

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  admin:
    api:
      restricted: 0

This is now deprecated, please refer to Configuration to find out about the new admin API mechanism.

HTTP Engine

proxy.config.http.server_ports

Scope:

CONFIG

Type:

STRING

Default:

8080 8080:ipv6

yaml:

records:
  http:
    server_ports: 8080 8080:ipv6

Ports used for proxying HTTP traffic.

This is a list, separated by space or comma, of port descriptors. Each descriptor is a sequence of keywords and values separated by colons. Not all keywords have values, those that do are specifically noted. Keywords with values can have an optional = character separating the keyword and value. The case of keywords is ignored. The order of keywords is irrelevant but unspecified results may occur if incompatible options are used (noted below). Options without values are idempotent. Options with values use the last (right most) value specified, except for ip-out as detailed later.

Quick reference chart:

Name	Note	Definition
number	Required	The local port.
blind		Blind (CONNECT) port.
compress	Not Implemented	Compressed.
ipv4	Default	Bind to IPv4 address family.
ipv6		Bind to IPv6 address family.
ip-in	Value	Local inbound IP address.
ip-out	Value	Local outbound IP address.
ip-resolve	Value	IP address resolution style.
proto	Value	List of supported session protocols.
pp		Enable Proxy Protocol.
ssl		SSL terminated.
quic		QUIC terminated.
tr-full		Fully transparent (inbound and outbound)
tr-in		Inbound transparent.
tr-out		Outbound transparent.
tr-pass		Pass through enabled.
mptcp		Multipath TCP.
allow-plain		Allow failback to non-TLS for TLS ports

number

Local IP port to bind. This is the port to which ATS clients will connect.

blind

Accept only the CONNECT method on this port.

Not compatible with: tr-in, ssl and quic.

compress

Compress the connection. Retained only by inertia, should be considered “not implemented”.

ipv4

Use IPv4. This is the default and is included primarily for completeness. This forced if the ip-in option is used with an IPv4 address.

ipv6

Use IPv6. This is forced if the ip-in option is used with an IPv6 address.

ssl

Require SSL termination for inbound connections. SSL must be configured for this option to provide a functional server port.

Not compatible with: blind and quic.

allow-plain allows a failback to non SSL for such ports.

quic

Require QUIC termination for inbound connections. SSL must be configured for this option to provide a functional server port. THIS IS EXPERIMENTAL SUPPORT AND NOT READY FOR PRODUCTION USE.

Not compatible with: blind and ssl.

proto

Specify the session level protocols supported. These should be separated by semi-colons. For TLS proxy ports the default value is all available protocols. For non-TLS proxy ports the default is HTTP only. HTTP/3 is only available on QUIC ports.

pp

Enables Proxy Protocol on the port. If Proxy Protocol is enabled on the port, all incoming requests must be prefaced with the PROXY header. See Proxy Protocol for more details on how to configure this option properly.

tr-full

Fully transparent. This is a convenience option and is identical to specifying both tr-in and tr-out.

Not compatible with: Any option not compatible with tr-in or tr-out.

tr-in

Inbound transparent. The proxy port will accept connections to any IP address on the port. To have IPv6 inbound transparent you must use this and the ipv6 option. This overrides proxy.config.incoming_ip_to_bind for this port.

Not compatible with: ip-in, blind

tr-out

Outbound transparent. If ATS connects to an origin server for a transaction on this port, it will use the client’s address as its local address. This overrides proxy.config.outgoing_ip_to_bind for this port.

Not compatible with: ip-out, ip-resolve

tr-pass

Transparent pass through. This option is useful only for inbound transparent proxy ports. If the parsing of the expected HTTP header fails, then the transaction is switched to a blind tunnel instead of generating an error response to the client. It effectively enables proxy.config.http.use_client_target_addr for the transaction as there is no other place to obtain the origin server address.

ip-in

Set the local IP address for the port. This is the address to which clients will connect. This forces the IP address family for the port. The ipv4 or ipv6 can be used but it is optional and is an error for it to disagree with the IP address family of this value. An IPv6 address must be enclosed in square brackets. If this option is omitted proxy.config.incoming_ip_to_bind is used.

Not compatible with: tr-in.

ip-out

Set the local IP address for outbound connections. This is the address used by ATS locally when it connects to an origin server for transactions on this port. If this is omitted proxy.config.outgoing_ip_to_bind is used.

This option can used multiple times, once for each IP address family. The address used is selected by the IP address family of the origin server address.

Not compatible with: tr-out.

ip-resolve

Set the host resolution style for transactions on this proxy port.

Not compatible with: tr-out - this option requires a value of client;none which is forced and should not be explicitly specified.

mptcp

Enable Multipath TCP on this proxy port.

Requires custom Linux kernel available at https://multipath-tcp.org.

allow-plain

For TLS ports, will fall back to non-TLS processing if the TLS handshake fails. Incompatible with quic ports.

Example

Listen on port 80 on any address for IPv4 and IPv6.:

80 80:ipv6

Example

Listen transparently on any IPv4 address on port 8080, and transparently on port 8080 on local address fc01:10:10:1::1 (which implies ipv6).:

IPv4:tr-FULL:8080 TR-full:IP-in=[fc02:10:10:1::1]:8080

Example

Listen on port 8080 for IPv6, fully transparent. Set up an SSL port on 443. These ports will use the IP address from proxy.config.incoming_ip_to_bind. Listen on IP address 192.168.17.1, port 80, IPv4, and connect to origin servers using the local address 10.10.10.1 for IPv4 and fc01:10:10:1::1 for IPv6.:

8080:ipv6:tr-full 443:ssl ip-in=192.168.17.1:80:ip-out=[fc01:10:10:1::1]:ip-out=10.10.10.1

Example

Listen on port 9090 for TLS enabled HTTP/2 or HTTP connections, accept no other session protocols.:

9090:proto=http2;http:ssl

Example

Listen on port 9090 for TLS disabled HTTP/2 and enabled HTTP connections, accept no other session protocols.:

9090:proto=http:ssl

Example

Listen on port 4433 for QUIC connections.:

4433:quic

proxy.config.http.connect_ports

Scope:

CONFIG

Type:

STRING

Default:

443

yaml:

records:
  http:
    connect_ports: '443'

The range of origin server ports that can be used for tunneling via CONNECT.

Traffic Server allows tunnels only to the specified ports. Values are space separated. Wildcards (*), ranges (0-1023), and mixes of each (e.g. 80 443 666-999) are supported.

Note

These are the ports on the origin server, not Traffic Server proxy ports.

proxy.config.http.forward_connect_method

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    forward_connect_method: 0

The default, Traffic Server behavior for handling a CONNECT method request is to establish a tunnel to the requested destination. This configuration alters the behavior so that Traffic Server forwards the CONNECT method to the next hop, and establishes the tunnel after receiving a positive response. This behavior is useful in a proxy hierarchy, and is equivalent to setting proxy.config.http.parent_proxy.disable_connect_tunneling to 0 when parent proxying is enabled.

proxy.config.http.insert_request_via_str

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    insert_request_via_str: 1

Set how the Via field is handled on a request to the origin server.

Value	Effect
0	Do not modify or set this Via header.
1	Add the basic protocol and proxy identifier.
2	Add basic transaction codes.
3	Add detailed transaction codes.
4	Add full user agent connection protocol tags.

Note

The Via transaction codes can be decoded with the Via Decoder Ring.

proxy.config.http.request_via_str

Scope:

CONFIG

Type:

STRING

Default:

ApacheTrafficServer/${PACKAGE_VERSION}

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    request_via_str: ApacheTrafficServer/${PACKAGE_VERSION}

Set the server and version string in the Via request header to the origin server which is inserted when the value of proxy.config.http.insert_request_via_str is not 0. Note that the actual default value is defined with "ApacheTrafficServer/" PACKAGE_VERSION in a C++ source code, and you must write such as ApacheTrafficServer/6.0.0 if you really set a value with the version in records.yaml file. If you want to hide the version, you can set this value to ApacheTrafficServer.

proxy.config.http.insert_response_via_str

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    insert_response_via_str: 0

Set how the Via field is handled on the response to the client.

Value	Effect
0	Do not modify or set this Via header.
1	Add the basic protocol and proxy identifier.
2	Add basic transaction codes.
3	Add detailed transaction codes.
4	Add full upstream connection protocol tags.

Note

The Via transaction code can be decoded with the Via Decoder Ring.

proxy.config.http.response_via_str

Scope:

CONFIG

Type:

STRING

Default:

ApacheTrafficServer/${PACKAGE_VERSION}

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    response_via_str: ApacheTrafficServer/${PACKAGE_VERSION}

Set the server and version string in the Via response header to the client which is inserted when the value of proxy.config.http.insert_response_via_str is not 0. Note that the actual default value is defined with "ApacheTrafficServer/" PACKAGE_VERSION in a C++ source code, and you must write such as ApacheTrafficServer/6.0.0 if you really set a value with the version in records.yaml file. If you want to hide the version, you can set this value to ApacheTrafficServer.

proxy.config.http.send_100_continue_response

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http:
    send_100_continue_response: 0

You can specify one of the following:

Value	Description
0	Traffic Server will buffer the request until the post body has been received and then send the request to the origin server.
1	Immediately return a 100 Continue from Traffic Server without waiting for the post body.

proxy.config.http.response_server_enabled

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    response_server_enabled: 1

You can specify one of the following:

Value	Description
0	No Server header is added to the response.
1	The Server header is added according to proxy.config.http.response_server_str.
2	The Server header is added only if the response from origin does not have one already.

proxy.config.http.response_server_str

Scope:

CONFIG

Type:

STRING

Default:

ATS/${PACKAGE_VERSION}

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    response_server_str: ATS/${PACKAGE_VERSION}

The Server string that Traffic Server will insert in a response header (if requested, see above). Note that the actual default value is defined with "ATS/" PACKAGE_VERSION in the C++ source, and you must write such as ATS/6.0.0 if you really set a value with the version in records.yaml. If you want to hide the version, you can set this value to ATS.

proxy.config.http.insert_age_in_response

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    insert_age_in_response: 1

This option specifies whether Traffic Server should insert an Age header in the response. The value is the cache’s estimate of the amount of time since the response was generated or revalidated by the origin server.

Value	Description
0	No Age header is added.
1	Age header is added.

proxy.config.http.chunking_enabled

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    chunking_enabled: 1

Specifies whether Traffic Server can generate a chunked response:

Value	Description
0	Never respond with chunked encoding.
1	Always respond with chunked encoding.
2	Generate a chunked response if the origin server has previously returned HTTP/1.1.
3	Generate a chunked response if the client request is HTTP/1.1 and the origin server has previously returned HTTP/1.1.

proxy.config.http.chunking.size

Scope:

CONFIG

Type:

INT

Default:

4096

Overridable:

Yes

yaml:

records:
  http:
    chunking:
      size: 4096

If chunked transfer encoding is enabled with proxy.config.http.chunking_enabled, and the conditions specified by that option’s setting are met by the current request, this option determines the size of the chunks, in bytes, to use when sending content to an HTTP/1.1 client.

proxy.config.http.drop_chunked_trailers

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    drop_chunked_trailers: 1

Specifies whether Traffic Server should drop chunked trailers. If enabled (1), Traffic Server will drop any chunked trailers in a Transfer-Encoded: chunked request or response body. If disabled (0), Traffic Server will pass the chunked trailers unmodified to the receiving peer. See RFC 9112, section 7.1.2 for details about chunked trailers. By default, this option is enabled and therefore Traffic Server will drop chunked trailers.

proxy.config.http.send_http11_requests

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    send_http11_requests: 1

Specifies when and how Traffic Server uses HTTP/1.1 to communicate with the origin server.

Value	Description
0	Never use HTTP/1.1.
1	Always use HTTP/1.1.
2	Use HTTP/1.1 with origin connections only if the server has previously returned HTTP/1.1.
3	If the client request is HTTP/1.1 and the origin server has previously returned HTTP/1.1, then use HTTP/1.1 for origin server connections.

Note

If proxy.config.http.use_client_target_addr is set to 1, then options 2 and 3 for this configuration variable cause the proxy to use the client HTTP version for upstream requests.

proxy.config.http.auth_server_session_private

Scope:

CONFIG

Type:

INT

Default:

1

Overridable:

Yes

yaml:

records:
  http:
    auth_server_session_private: 1

If enabled (1) anytime a request contains a Authorization, Proxy-Authorization, or Www-Authenticate header the connection will be closed and not reused. This marks the connection as private. When disabled (0) the connection will be available for reuse.

proxy.config.http.server_session_sharing.match

Scope:

CONFIG

Type:

STRING

Default:

both

Overridable:

Yes

yaml:

records:
  http:
    server_session_sharing:
      match: both

Enable and set the ability to re-use server connections across client connections. Multiple values can be specified when separated by commas with no white spaces. Valid values are:

Value	Description
none	Do not match and do not re-use server sessions.
ip	Re-use server sessions, checking only that the IP address and port of the origin server matches.
host	Re-use server sessions, checking that the fully qualified domain name matches. In addition, if the session uses TLS, it also checks that the current transaction’s host header value matches the session’s SNI.
both	Equivalent to host,ip.
hostonly	Check that the fully qualified domain name matches.
sni	Check that the SNI of the session matches the SNI that would be used to create a new session. Only applicable for TLS sessions.
cert	Check that the certificate file name used for the server session matches the certificate file name that would be used for the new server session. Only applicable for TLS sessions.

The setting must contain at least one of ip, host, hostonly or both for session reuse to operate. The other values may be used for greater control with TLS session reuse.

Note

Server sessions to different upstream ports never match even if the FQDN and IP address match.

Note

Upstream session tracking uses a similar set of options for matching sessions, but is set independently from session sharing.

proxy.config.http.server_session_sharing.pool

Scope:

CONFIG

Type:

STRING

Default:

thread

yaml:

records:
  http:
    server_session_sharing:
      pool: thread

Control the scope of server session re-use if it is enabled by proxy.config.http.server_session_sharing.match. Valid values are:

Value	Description
global	Re-use sessions from a global pool of all server sessions.
thread	Re-use sessions from a per-thread pool.
hybrid	Try to work as a global pool, but release server sessions to the per-thread pool if there is lock contention on the global pool.
global_locked	Similar to global, except that the session pool is managed by a blocking mutex.

Setting proxy.config.http.server_session_sharing.pool to global can reduce the number of connections to origin for some traffic loads. However, if many execute threads are active, the thread contention on the global pool can reduce the lifetime of connections to origin and reduce effective origin connection reuse.

For a hybrid pool, the operation starts as the global pool, but sessons are returned to the local thread pool if the global pool lock is not acquired rather than just closing the origin connection as is the case in standard global mode.

For a global_locked pool connections are managed by a blocking mutex instead of the normal try mutex. Under extreme transaction loads the connection pool starvation may result in most transactions bypassing the connection pool resulting in runaway upstream connections. This option will avoid this condition at the cost of latency and ttfb (time to first byte) performance).

proxy.config.http.attach_server_session_to_client

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  http:
    attach_server_session_to_client: 0

Control the re-use of an server session by a user agent (client) session. Currently only applies to user agents using HTTP/1.0 or HTTP/1.1. For other HTTP versions, the origin connection is always returned to the session sharing pool or closed.

If a user agent performs more than one HTTP transaction on its connection to Traffic Server a server session must be obtained for the second (and subsequent) transaction as for the first. This settings affects how that server session is selected.

If this setting is 0 then after the first transaction the server session for that transaction is released to the server pool (if any). When a server session is needed for subsequent transactions one is selected from the server pool or created if there is no suitable server session in the pool.

If this setting is not 0 then the current server session for the user agent session is “sticky”. It will be preferred to any other server session (either from the pool or newly created). The server session will be detached from the user agent session only if it cannot be used for the transaction. This is determined by the proxy.config.http.server_session_sharing.match value. If the server session matches the next transaction according to this setting then it will be used, otherwise it will be released to the pool and a different session selected or created.

proxy.config.http.max_proxy_cycles

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  http:
    max_proxy_cycles: 0

Control the proxy cycle detection function in the following manner –

If this setting is 0, then next hop is self IP address and port detection is active.

In addition, the proxy cycle detection using the Via string will declare a cycle if the current cache appears one or more times in the Via string, i.e, > 0.

If this setting is 1 or more (N), then next hop is self IP address and port detection is inactive.

In addition, the proxy cycle detection using the Via string will declare a cycle if the current cache appears more than N times in the Via string, i.e., > N.

Examples:

If the setting is 0, then the second time a request enters a cache it will have its own machine identifier in the Via string once, so a cycle will be detected. So no cycles are allowed.

If the setting is 1, then the third time a request enters a cache it will have its own machine identifier in the Via string twice, so a cycle will be detected. So one cycle is allowed. The first cycle with two visits to the cache and one instance in the Via string is allowed. The second cycle with three visits to the cache and two instances in the Via string is not allowed.

This setting allows an edge cache peering arrangement where an edge cache may forward a request to a peer edge cache (possibly itself) a limited of times (usually once). Infinite loops are still detected when the cycle allowance is exceeded.

proxy.config.http.use_client_target_addr

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  http:
    use_client_target_addr: 0

For fully transparent ports use the same origin server address as the client.

This option causes Traffic Server to avoid where possible doing DNS lookups in forward transparent proxy mode. The option is only effective if the following three conditions are true:

• Traffic Server is in forward proxy mode.

• The proxy port is inbound transparent.

• The target URL has not been modified by either remapping or a plugin.

If any of these conditions are not true, then normal DNS processing is done for the connection.

There are three valid values.

Value	Description
0	Disables the feature.
1	Enables the feature with address verification. The proxy does the regular DNS processing. If the client-specified origin address is not in the set of addresses found by the proxy, the request continues to the client specified address, but the result is not cached.
2	Enables the feature with no address verification. No DNS processing is performed. The result is cached (if allowed otherwise). This option is vulnerable to cache poisoning if an incorrect Host header is specified, so this option should be used with extreme caution if HTTP caching is enabled. See bug TS-2954 for details.

If all of these conditions are met, then the origin server IP address is retrieved from the original client connection, rather than through HostDB or DNS lookup. In effect, client DNS resolution is used instead of Traffic Server DNS.

This can be used to be a little more efficient (looking up the target once by the client rather than by both the client and Traffic Server) but the primary use is when client DNS resolution can differ from that of Traffic Server. Two known uses cases are:

1. Embedded IP addresses in a protocol with DNS load sharing. In this case, even though Traffic Server and the client both make the same request to the same DNS resolver chain, they may get different origin server addresses. If the address is embedded in the protocol then the overall exchange will fail. One current example is Microsoft Windows update, which presumably embeds the address as a security measure.

2. The client has access to local DNS zone information which is not available to Traffic Server. There are corporate nets with local DNS information for internal servers which, by design, is not propagated outside the core corporate network. Depending a network topology it can be the case that Traffic Server can access the servers by IP address but cannot resolve such addresses by name. In such as case the client supplied target address must be used.

This solution must be considered interim. In the longer term, it should be possible to arrange for much finer grained control of DNS lookup so that wildcard domain can be set to use Traffic Server or client resolution. In both known use cases, marking specific domains as client determined (rather than a single global switch) would suffice. It is possible to do this crudely with this flag by enabling it and then use identity URL mappings to re-disable it for specific domains.

proxy.config.http.keep_alive_enabled_in

Scope:

CONFIG

Type:

INT

Default:

1

Overridable:

Yes

yaml:

records:
  http:
    keep_alive_enabled_in: 1

Enables (1) or disables (0) incoming keep-alive connections.

proxy.config.http.keep_alive_enabled_out

Scope:

CONFIG

Type:

INT

Default:

1

Overridable:

Yes

yaml:

records:
  http:
    keep_alive_enabled_out: 1

Enables (1) or disables (0) outgoing keep-alive connections.

Note

Enabling keep-alive does not automatically enable purging of keep-alive requests when nearing the connection limit, that is controlled by proxy.config.http.server_max_connections.

proxy.config.http.keep_alive_post_out

Scope:

CONFIG

Type:

INT

Default:

1

Overridable:

Yes

yaml:

records:
  http:
    keep_alive_post_out: 1

Controls whether new POST requests re-use keep-alive sessions (1) or create new connections per request (0).

proxy.config.http.disallow_post_100_continue

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  http:
    disallow_post_100_continue: 0

Allows you to return a 405 Method Not Supported with Posts also containing an Expect: 100-continue.

When a Post w/ Expect: 100-continue is blocked the stat proxy.process.http.disallowed_post_100_continue will be incremented.

proxy.config.http.default_buffer_size

Scope:

CONFIG

Type:

INT

Default:

8

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    default_buffer_size: 8

Configures the default buffer size, in bytes, to allocate for incoming request bodies which lack a Content-length header.

proxy.config.http.default_buffer_water_mark

Scope:

CONFIG

Type:

INT

Default:

32768

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    default_buffer_water_mark: 32768

Number of bytes Traffic Server is allowed to read ahead of the client from the origin. Note that when Read While Write settings are in place, this setting will apply to the first client to request the object, regardless if subsequent, simultaneous clients of that object can read faster. The buffered bytes will consume memory while waiting for the client to consume them.

While this setting is reloadable, dramatic changes can cause bigger memory usage than expected and is thus not recommended.

proxy.config.http.request_buffer_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  http:
    request_buffer_enabled: 0

This enables buffering the content for incoming POST requests. If enabled no outbound connection is made until the entire POST request has been buffered. If enabled, proxy.config.http.post_copy_size needs to be set to the maximum of the post body size allowed, otherwise, the post would fail.

proxy.config.http.request_line_max_size

Scope:

CONFIG

Type:

INT

Default:

65535

Reloadable:

Yes

yaml:

records:
  http:
    request_line_max_size: 65535

Controls the maximum size, in bytes, of an HTTP Request Line in requests. Requests with a request line exceeding this size will be treated as invalid and rejected by the proxy. Note that the HTTP request line typically includes HTTP method, request target and HTTP version string except when the request is made using absolute URI in which case the request line may also include the request scheme and domain name.

proxy.config.http.header_field_max_size

Scope:

CONFIG

Type:

INT

Default:

32768

Reloadable:

Yes

yaml:

records:
  http:
    header_field_max_size: 32768

Controls the maximum size, in bytes, of an HTTP header field in requests. Headers in a request with the sum of their name and value that exceed this size will cause the entire request to be treated as invalid and rejected by the proxy.

proxy.config.http.request_header_max_size

Scope:

CONFIG

Type:

INT

Default:

32768

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    request_header_max_size: 32768

Controls the maximum size, in bytes, of an HTTP header in requests. Headers in a request which exceed this size will cause the entire request to be treated as invalid and rejected by the proxy.

proxy.config.http.response_header_max_size

Scope:

CONFIG

Type:

INT

Default:

32768

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    response_header_max_size: 32768

Controls the maximum size, in bytes, of headers in HTTP responses from the proxy. Any responses with a header exceeding this limit will be treated as invalid and a client error will be returned instead.

proxy.config.http.global_user_agent_header

Scope:

CONFIG

Type:

STRING

Default:

null

Overridable:

Yes

yaml:

records:
  http:
    global_user_agent_header: 'null'

An arbitrary string value that, if set, will be used to replace any request User-Agent header.

proxy.config.http.strict_uri_parsing

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  http:
    strict_uri_parsing: 2

Takes a value between 0 and 2. 0 disables strict_uri_parsing. Any character can appears in the URI. 1 causes Traffic Server to return 400 Bad Request if client’s request URI includes character which is not RFC 3986 compliant. 2 directs Traffic Server to reject the clients request if it contains whitespace or non-printable characters.

proxy.config.http.errors.log_error_pages

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  http:
    errors:
      log_error_pages: 1

Enables (1) or disables (0) the logging of responses to bad requests to the error logging destination. Disabling this option prevents error responses (such as 403s) from appearing in the error logs. Any HTTP response status codes equal to, or higher, than the minimum code defined by TS_HTTP_STATUS_BAD_REQUEST are affected by this setting.

Parent Proxy Configuration

proxy.config.http.parent_proxy.retry_time

Scope:

CONFIG

Type:

INT

Default:

300

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    parent_proxy:
      retry_time: 300

The amount of time allowed between connection retries to a parent cache that is unavailable.

proxy.config.http.parent_proxy.max_trans_retries

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  http:
    parent_proxy:
      max_trans_retries: 2

Limits the number of simultaneous transactions that may retry a parent once the parents retry_time has expired.

proxy.config.http.parent_proxy.fail_threshold

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    parent_proxy:
      fail_threshold: 10

The number of times the connection to the parent cache can fail before Traffic Server considers the parent unavailable.

proxy.config.http.parent_proxy.total_connect_attempts

Scope:

CONFIG

Type:

INT

Default:

4

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    parent_proxy:
      total_connect_attempts: 4

The total number of connection attempts for a specific transaction allowed to a parent cache before Traffic Server bypasses the parent or fails the request (depending on the go_direct option in the parent.config file). The number of parents tried is proxy.config.http.parent_proxy.fail_threshold / proxy.config.http.parent_proxy.total_connect_attempts

proxy.config.http.parent_proxy.per_parent_connect_attempts

Scope:

CONFIG

Type:

INT

Default:

2

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    parent_proxy:
      per_parent_connect_attempts: 2

The total number of connection attempts allowed per parent for a specific transaction, if multiple parents are used.

proxy.config.http.parent_proxy.mark_down_hostdb

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    parent_proxy:
      mark_down_hostdb: 0

Enables (1) or disables (0) marking parent proxies down in hostdb when a connection error is detected. Normally parent selection manages parent proxies and will mark them as unavailable as needed. But when parents are defined in dns with multiple ip addresses, it may be useful to mark the failing ip down in hostdb. In this case you would enable these updates.

proxy.config.http.forward.proxy_auth_to_parent

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    forward:
      proxy_auth_to_parent: 0

Configures Traffic Server to send proxy authentication headers on to the parent cache.

proxy.config.http.no_dns_just_forward_to_parent

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    no_dns_just_forward_to_parent: 0

Don’t try to resolve DNS, forward all DNS requests to the parent. This is off (0) by default.

proxy.config.http.parent_proxy.disable_connect_tunneling

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  http:
    parent_proxy:
      disable_connect_tunneling: 0

proxy.config.http.parent_proxy.self_detect

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  http:
    parent_proxy:
      self_detect: 2

For each host that has been specified in a parent or secondary_parent list in the parent.config file, determine if the host is the same as the current host. Obvious examples include localhost and 127.0.0.1. If a match is found, take an action depending upon the value below.

Value	Description
0	Disables the feature by not checking for matches.
1	Remove the matching host from the list.
2	Mark the host down. This is the default.

proxy.config.http.parent_proxy.enable_parent_timeout_markdowns

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    parent_proxy:
      enable_parent_timeout_markdowns: 0

Enables (1) or disables (0) parent proxy mark downs due to inactivity timeouts. By default parent proxies are not marked down due to inactivity timeouts, the transaction will retry using another parent instead. The default for this configuration keeps this behavior and is disabled (0). This setting is overridable using one of the two plugins header_rewrite or conf_remap to enable inactivity timeout markdowns and should be done so rather than enabling this globally. This setting should not be used in conjunction with proxy.config.http.parent_proxy.disable_parent_markdowns

proxy.config.http.parent_proxy.disable_parent_markdowns

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    parent_proxy:
      disable_parent_markdowns: 0

Enables (1) or disables (0) parent proxy markdowns. This is useful if parent entries in a parent.config line are VIP’s and one doesn’t wish to mark down a VIP which may have several origin or parent proxies behind the load balancer. This setting is overridable using one of the header_rewrite or the conf_remap plugins to override the default setting and this method should be used rather than disabling markdowns globally. This setting should not be used in conjunction with proxy.config.http.parent_proxy.enable_parent_timeout_markdowns

HTTP Connection Timeouts

proxy.config.http.keep_alive_no_activity_timeout_in

Scope:

CONFIG

Type:

INT

Default:

120

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    keep_alive_no_activity_timeout_in: 120

Specifies how long Traffic Server keeps connections to clients open for a subsequent request after a transaction ends. A value of 0 will set proxy.config.net.default_inactivity_timeout as the timeout.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.keep_alive_no_activity_timeout_out

Scope:

CONFIG

Type:

INT

Default:

120

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    keep_alive_no_activity_timeout_out: 120

Specifies how long Traffic Server keeps connections to origin servers open for a subsequent transfer of data after a transaction ends. A value of 0 will set proxy.config.net.default_inactivity_timeout as the timeout.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.transaction_no_activity_timeout_in

Scope:

CONFIG

Type:

INT

Default:

30

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    transaction_no_activity_timeout_in: 30

Specifies how long Traffic Server keeps connections to clients open if a transaction stalls.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.transaction_no_activity_timeout_out

Scope:

CONFIG

Type:

INT

Default:

30

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    transaction_no_activity_timeout_out: 30

Specifies how long Traffic Server keeps connections to origin servers open if the transaction stalls.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.websocket.no_activity_timeout

Scope:

CONFIG

Type:

INT

Default:

600

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  websocket:
    no_activity_timeout: 600

Specifies how long Traffic Server keeps connections open if a websocket stalls.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.websocket.active_timeout

Scope:

CONFIG

Type:

INT

Default:

3600

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  websocket:
    active_timeout: 3600

The maximum amount of time Traffic Server keeps websocket connections open.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.transaction_active_timeout_in

Scope:

CONFIG

Type:

INT

Default:

900

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    transaction_active_timeout_in: 900

The maximum amount of time Traffic Server can remain connected to a client. If the transfer to the client is not complete before this timeout expires, then Traffic Server closes the connection.

The value of 0 specifies that there is no timeout.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.transaction_active_timeout_out

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    transaction_active_timeout_out: 0

The maximum amount of time Traffic Server waits for fulfillment of a connection request to an origin server. If Traffic Server does not complete the transfer to the origin server before this timeout expires, then Traffic Server terminates the connection request.

The default value of 0 specifies that there is no timeout.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.accept_no_activity_timeout

Scope:

CONFIG

Type:

INT

Default:

120

Reloadable:

Yes

yaml:

records:
  http:
    accept_no_activity_timeout: 120

The timeout interval in seconds before Traffic Server closes a connection that has no activity.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.background_fill_active_timeout

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    background_fill_active_timeout: 0

Specifies how long Traffic Server continues a background fill before giving up and dropping the origin server connection.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.background_fill_completed_threshold

Scope:

CONFIG

Type:

FLOAT

Default:

0.0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    background_fill_completed_threshold: 0.0

The proportion of total document size already transferred when a client aborts at which the proxy continues fetching the document from the origin server to get it into the cache (a background fill).

HTTP Redirection

proxy.config.http.number_of_redirections

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    number_of_redirections: 0

This setting determines the maximum number of times Trafficserver does a redirect follow location on receiving a 3XX Redirect response for a given client request.

Note

When proxy.config.http.number_of_redirections is set to a positive value and Traffic Server has previously cached a 3XX Redirect response, the cached response will continue to be refreshed and returned until the response is no longer in the cache.

Note

In previous versions proxy.config.http.redirection_enabled had to be set to 1 before this setting was evaluated. Now setting proxy.config.http.number_of_redirections to a value greater than zero is sufficient to cause Traffic Server to follow redirects.

proxy.config.http.redirect_host_no_port

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  http:
    redirect_host_no_port: 1

This setting enables Trafficserver to not include the port in the Host header in the redirect follow request for default/standard ports (e.g. 80 for HTTP and 443 for HTTPS). Note that the port is still included in the Host header if it’s non-default.

proxy.config.http.redirect_use_orig_cache_key

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    redirect_use_orig_cache_key: 0

This setting enables Trafficserver to allow using original request cache key (for example, set using a TS API) during a 3xx redirect follow. The default behavior (0) is to use the URL specified by Location header in the 3xx response as the cache key.

proxy.config.http.post_copy_size

Scope:

CONFIG

Type:

INT

Default:

2048

Reloadable:

Yes

yaml:

records:
  http:
    post_copy_size: 2048

This setting determines the maximum size in bytes of uploaded content to be buffered for HTTP methods such as POST and PUT.

proxy.config.http.redirect.actions

Scope:

CONFIG

Type:

STRING

Default:

routable:follow

Reloadable:

Yes

yaml:

records:
  http:
    redirect:
      actions: routable:follow

This setting determines how redirects should be handled. The setting consists of a comma-separated list of key-value pairs, where the keys are named IP address ranges and the values are actions.

The following are valid keys:

Key	Description
self	Addresses of the host’s interfaces
loopback	IPv4 127.0.0.0/8 and IPv6 ::1
private	IPv4 10.0.0.0/8 100.64.0.0/10 172.16.0.0/12 192.168.0.0/16 and IPv6 fc00::/7
multicast	IPv4 224.0.0.0/4 and IPv6 ff00::/8
linklocal	IPv4 169.254.0.0/16 and IPv6 fe80::/10
routable	All publicly routable addresses
default	All address ranges not configured specifically

The following are valid values:

Value	Description
return	Do not process the redirect, send it as the proxy response.
reject	Do not process the redirect, send a 403 as the proxy response.
follow	Internally follow the redirect up to proxy.config.http.number_of_redirections. Use this setting with caution!

Warning

Following a redirect to other than routable addresses can be dangerous, as it allows the controller of an origin to arrange a probe the Traffic Server host. Enabling these redirects makes Traffic Server open to third party attacks and probing and therefore should be considered only in known safe environments.

For example, a setting of loopback:reject,private:reject,routable:follow,default:return would send 403 as the proxy response to loopback and private addresses, routable addresses would be followed up to proxy.config.http.number_of_redirections, and redirects to all other ranges will be sent as the proxy response.

The action for self has the highest priority when an address would match multiple keys, and the action for default has the lowest priority. Other keys represent disjoint sets of addresses that will not conflict. If duplicate keys are present in the setting, the right-most key-value pair is used.

The default value is routable:follow, which means “follow routable redirects, return all other redirects”. Note that proxy.config.http.number_of_redirections must be positive also, otherwise redirects will be returned rather than followed.

Origin Server Connect Attempts

proxy.config.http.connect_attempts_max_retries

Scope:

CONFIG

Type:

INT

Default:

3

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    connect_attempts_max_retries: 3

The maximum number of connection retries Traffic Server can make when the origin server is not responding. Each retry attempt lasts for proxy.config.http.connect_attempts_timeout seconds. Once the maximum number of retries is reached, the origin is marked down (as controlled by proxy.config.http.connect.down.policy. After this, the setting proxy.config.http.connect_attempts_max_retries_down_server is used to limit the number of retry attempts to the known down origin.

proxy.config.http.connect_attempts_max_retries_down_server

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    connect_attempts_max_retries_down_server: 1

Maximum number of connection attempts Traffic Server can make while an origin is marked down per request. Typically this value is smaller than proxy.config.http.connect_attempts_max_retries so an error is returned to the client faster and also to reduce the load on the down origin. The timeout interval proxy.config.http.connect_attempts_timeout in seconds is used with this setting.

proxy.config.http.connect.down.policy

Scope:

CONFIG

Type:

INT

Default:

2

Overridable:

Yes

yaml:

records:
  http:
    connect:
      down:
        policy: 2

Controls what origin server connection failures contribute to marking a server down. When set to 2, any connection failure during the TCP and TLS handshakes will contribute to marking the server down. When set to 1, only TCP handshake failures will contribute to marking a server down. When set to 0, no connection failures will be used towards marking a server down.

proxy.config.http.server_max_connections

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http:
    server_max_connections: 0

Limits the number of socket connections across all origin servers to the value specified. To disable, set to zero (0).

This value is used in determining when and if to prune active origin sessions. Without this value set, connections to origins can consume all the way up to proxy.config.net.connections_throttle connections, which in turn can starve incoming requests from available connections.

proxy.config.http.per_server.connection.max

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    per_server:
      connection:
        max: 0

Set a limit for the number of concurrent connections to an upstream server group. A value of 0 disables checking. If a transaction attempts to connect to a group which already has the maximum number of concurrent connections a 503 (HTTP_STATUS_SERVICE_UNAVAILABLE) error response is sent to the user agent. To configure

Upstream server group definition

See proxy.config.http.per_server.connection.match.

Frequency of alerts

See proxy.config.http.per_server.connection.alert_delay.

proxy.config.http.per_server.connection.match

Scope:

CONFIG

Type:

STRING

Default:

both

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    per_server:
      connection:
        match: both

Control the definition of an upstream server group for proxy.config.http.per_server.connection.max. This must be one of the following keywords.

ip

Group by IP address. Each IP address is a group.

port

Group by IP address and port. Each distinct IP address and port pair is a group.

host

Group by host name. The host name is the post remap FQDN used to resolve the upstream address.

both

Group by IP address, port, and host name. Each distinct combination is a group.

To disable upstream server grouping, set proxy.config.http.per_server.connection.max to 0.

Note

This setting is independent of the setting for upstream session sharing matching.

proxy.config.http.per_server.connection.alert_delay

Scope:

CONFIG

Type:

INT

Default:

60

Units:

seconds

Reloadable:

Yes

yaml:

records:
  http:
    per_server:
      connection:
        alert_delay: 60

Throttle alerts per upstream server group to be no more often than this many seconds. Summary data is provided per alert to allow log scrubbing to generate accurate data.

proxy.config.http.per_server.connection.min

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    per_server:
      connection:
        min: 0

Set a target for the minimum number of active connections to an upstream server group. When an outbound connection is in keep alive state and the inactivity timer expires, if there are fewer than this many connections in the group a new connection the timer is reset instead of closing the connection. Useful when the origin supports keep-alive, removing the time needed to set up a new connection from the next request at the expense of added (inactive) connections.

proxy.config.http.connect_attempts_rr_retries

Scope:

CONFIG

Type:

INT

Default:

3

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    connect_attempts_rr_retries: 3

The maximum number of failed connection attempts allowed before a round-robin entry is marked as ‘down’ if a server has round-robin DNS entries.

proxy.config.http.connect_attempts_timeout

Scope:

CONFIG

Type:

INT

Default:

30

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    connect_attempts_timeout: 30

The timeout value (in seconds) for time to set up a connection to the origin. After the connection is established the value of proxy.config.http.transaction_no_activity_timeout_out is used to established timeouts on the data over the connection.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.http.post.check.content_length.enabled

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  http:
    post:
      check:
        content_length:
          enabled: 1

Enables (1) or disables (0) checking the Content-Length: Header for a POST request.

proxy.config.http.down_server.cache_time

Scope:

CONFIG

Type:

INT

Default:

60

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    down_server:
      cache_time: 60

Specifies how long (in seconds) Traffic Server remembers that an origin server was unreachable.

proxy.config.http.uncacheable_requests_bypass_parent

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    uncacheable_requests_bypass_parent: 1

When enabled (1), Traffic Server bypasses the parent proxy for a request that is not cacheable.

Congestion Control

proxy.config.http.flow_control.enabled

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  http:
    flow_control:
      enabled: 0

Transaction buffering / flow control is enabled if this is set to a non-zero value. Otherwise no flow control is done.

proxy.config.http.flow_control.high_water

Scope:

CONFIG

Type:

INT

Default:

0

Units:

bytes

Overridable:

Yes

yaml:

records:
  http:
    flow_control:
      high_water: 0

The high water mark for transaction buffer control. External source I/O is halted when the total buffer space in use by the transaction exceeds this value.

proxy.config.http.flow_control.low_water

Scope:

CONFIG

Type:

INT

Default:

0

Units:

bytes

Overridable:

Yes

yaml:

records:
  http:
    flow_control:
      low_water: 0

The low water mark for transaction buffer control. External source I/O is resumed when the total buffer space in use by the transaction is no more than this value.

proxy.config.http.websocket.max_number_of_connections

Scope:

CONFIG

Type:

INT

Default:

-1

Reloadable:

Yes

yaml:

records:
  http:
    websocket:
      max_number_of_connections: -1

When enabled >= (0), Traffic Server will enforce a maximum number of simultaneous websocket connections.

Negative Response Caching

proxy.config.http.negative_caching_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    negative_caching_enabled: 0

When enabled (1), Traffic Server caches negative responses (such as 404 Not Found) when a requested page does not exist. The next time a client requests the same page, Traffic Server serves the negative response directly from cache.

When disabled (0), Traffic Server will only cache the response if the response has Cache-Control headers.

The following negative responses are cached by Traffic Server by default:

HTTP Response Code	Description
204	No Content
305	Use Proxy
403	Forbidden
404	Not Found
414	URI Too Long
500	Internal Server Error
501	Not Implemented
502	Bad Gateway
503	Service Unavailable
504	Gateway Timeout

The cache lifetime for objects cached from this setting is controlled via proxy.config.http.negative_caching_lifetime.

proxy.config.http.negative_caching_lifetime

Scope:

CONFIG

Type:

INT

Default:

1800

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    negative_caching_lifetime: 1800

How long (in seconds) Traffic Server keeps the negative responses valid in cache. This value only affects negative responses that do NOT have explicit Expires: or Cache-Control: lifetimes set by the server.

proxy.config.http.negative_caching_list

Scope:

CONFIG

Type:

STRING

Default:

204 305 403 404 414 500 501 502 503 504

Reloadable:

Yes

yaml:

records:
  http:
    negative_caching_list: 204 305 403 404 414 500 501 502 503 504

The HTTP status code for negative caching. Default values are mentioned above. The unwanted status codes can be taken out from the list. Other status codes can be added. The variable is a list but parsed as STRING.

proxy.config.http.negative_revalidating_enabled

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    negative_revalidating_enabled: 1

Negative revalidating allows Traffic Server to return stale content if revalidation to the origin fails due to network or HTTP errors. If it is enabled, rather than caching the negative response, the current stale content is preserved and served. Note this is considered only on a revalidation of already cached content. A revalidation failure means a connection failure or a 50x response code. When considering replying with a stale response in these negative revalidating circumstances, Traffic Server will respect the proxy.config.http.cache.max_stale_age configuration and will not use a cached response older than max_stale_age seconds.

A value of 0 disables serving stale content and a value of 1 enables keeping and serving stale content if revalidation fails.

proxy.config.http.negative_revalidating_lifetime

Scope:

CONFIG

Type:

INT

Default:

1800

yaml:

records:
  http:
    negative_revalidating_lifetime: 1800

When replying with a stale cached response in negative revalidating circumstances (see proxy.config.http.negative_revalidating_enabled), Traffic Server includes an Expires: HTTP header field in the cached response with a future time so that upstream caches will not try to revalidate their respective stale objects. This configuration specifies how many seconds in the future Traffic Server will calculate the value of this inserted Expires: header field.

There is a limitation to this method to be aware of: per specification (see IETF RFC 7234, section 4.2.1), Cache-Control: response directives take precedence over the Expires: header field when determining object freshness. Thus if the cached response contains either a max-age or an s-maxage Cache-Control: response directive, then these directives would take precedence for the upstream caches over the inserted Expires: field, rendering the Expires: header ineffective in specifying the configured freshness lifetime.

Finally, be aware that the only way this configuration is used is as input into calculating the value of these inserted Expires: header fields. This configuration does not direct Traffic Server behavior with regard to whether it considers a stale object to be fresh enough to serve out of cache when revalidation fails. As mentioned above in proxy.config.http.negative_revalidating_enabled, proxy.config.http.cache.max_stale_age is used for that determination.

This configuration defaults to 1,800 seconds (30 minutes).

Proxy User Variables

proxy.config.http.anonymize_remove_from

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    anonymize_remove_from: 0

When enabled (1), Traffic Server removes the From header to protect the privacy of your users.

proxy.config.http.anonymize_remove_referer

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http:
    anonymize_remove_referer: 0

When enabled (1), Traffic Server removes the Referrer header to protect the privacy of your site and users.

proxy.config.http.anonymize_remove_user_agent

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    anonymize_remove_user_agent: 0

When enabled (1), Traffic Server removes the User-agent header to protect the privacy of your site and users.

proxy.config.http.anonymize_remove_cookie

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    anonymize_remove_cookie: 0

When enabled (1), Traffic Server removes the Cookie header to protect the privacy of your site and users.

proxy.config.http.anonymize_remove_client_ip

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    anonymize_remove_client_ip: 0

When enabled (1), Traffic Server removes Client-IP headers for more privacy.

proxy.config.http.insert_client_ip

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    insert_client_ip: 1

Specifies whether Traffic Server inserts Client-IP headers to retain the client IP address:

Value	Description
0	Don’t insert the Client-ip header
1	Insert the Client-ip header, but only if the UA did not send one
2	Always insert the Client-ip header

proxy.config.http.anonymize_other_header_list

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  http:
    anonymize_other_header_list: null

Comma separated list of headers Traffic Server should remove from outgoing requests.

proxy.config.http.insert_squid_x_forwarded_for

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    insert_squid_x_forwarded_for: 1

When enabled (1), Traffic Server adds the client IP address to the X-Forwarded-For header.

proxy.config.http.insert_forwarded

Scope:

CONFIG

Type:

STRING

Default:

none

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    insert_forwarded: none

The default value (none) means that Traffic Server does not insert or append information to any Forwarded header (described in IETF RFC 7239) in the request message. To put information into a Forwarded header in the request, the value of this variable must be a list of the Forwarded parameters to be inserted.

Parameter	Value of parameter place in outgoing Forwarded header
for	Client IP address
by=ip	Proxy IP address
by=unknown	The literal string unknown
by=servername	Proxy server name
by=uuid	Server UUID prefixed with _
proto	Protocol of incoming request
host	The host specified in the incoming request
connection=compact	Connection with basic transaction codes.
connection=std	Connection with detailed transaction codes.
connection=full	Full user agent connection protocol tags

Each parameter in the list must be separated by | or :. For example, for|by=uuid|proto is a valid value for this variable. Note that the connection parameter is a non-standard extension to RFC 7239. Also note that, while Traffic Server allows multiple by parameters for the same proxy, this is prohibited by RFC 7239. Currently, for the host parameter to provide the original host from the incoming client request, proxy.config.url_remap.pristine_host_hdr must be enabled.

proxy.config.http.proxy_protocol_allowlist

Scope:

CONFIG

Type:

STRING

Default:

```<ip list>```

yaml:

records:
  http:
    proxy_protocol_allowlist: '```<ip list>```'

This defines a allowlist of server IPs that are trusted to provide connections with Proxy Protocol information. This is a comma delimited list of IP addresses. Addressed may be listed individually, in a range separated by a dash or by using CIDR notation.

Example Effect
10.0.2.123	A single IP Address.
10.0.3.1-10.0.3.254	A range of IP address.
10.0.4.0/24	A range of IP address specified by CIDR notation.

Important

If Proxy Protocol is enabled on the port, but this directive is not defined any server may initiate a connection with Proxy Protocol information. See proxy.config.http.server_ports for information on how to enable Proxy Protocol on a port.

See Proxy Protocol for more discussion on how Traffic Server transforms the Forwarded: header.

proxy.config.http.proxy_protocol_out

Scope:

CONFIG

Type:

INT

Default:

-1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    proxy_protocol_out: -1

Set the behavior of outbound PROXY Protocol.

Value	Description
-1	Disable (default)
0	Forward received PROXY protocol to the next hop
1	Send client information in PROXY protocol version 1
2	Send client information in PROXY protocol version 2

proxy.config.http.normalize_ae

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    normalize_ae: 1

Specifies normalization, if any, of Accept-Encoding: headers.

Value	Description
0	No normalization.
1	Accept-Encoding: gzip (if the header has gzip or x-gzip with any q) OR blank (for any header that does not include gzip)
2	Accept-Encoding: br if the header has br (with any q) ELSE normalize as for value 1
3	Accept-Encoding: br, gzip (if the header has br and gzip (with any q for either) then br, gzip) ELSE normalize as for value 2

This is useful for minimizing cached alternates of documents (e.g. gzip, deflate vs. deflate, gzip). Enabling this option is recommended if your origin servers use no encodings other than gzip or br (Brotli).

Security

proxy.config.http.push_method_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http:
    push_method_enabled: 0

Enables (1) or disables (0) the HTTP PUSH option, which allows you to deliver content directly to the cache without a user request.

Important

If you enable this option, then you must also specify a filtering rule in the ip_allow.yaml file to allow only certain machines to push content into the cache.

proxy.config.http.max_post_size

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http:
    max_post_size: 0

This feature is disabled by default with a value of (0), any positive value will limit the size of post bodies. If a request is received with a post body larger than this limit the response will be terminated with 413 - Request Entity Too Large and logged accordingly.

proxy.config.http.allow_multi_range

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    allow_multi_range: 0

This option allows the administrator to configure different behavior and handling of requests with multiple ranges in the Range header.

Value	Description
0	Do not allow multiple ranges, effectively ignoring the Range header
1	Allows multiple ranges. This can be potentially dangerous since well formed requests can cause excessive resource consumption on the server.
2	Similar to 0, except return a 416 error code and no response body.

proxy.config.http.host_sni_policy

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  http:
    host_sni_policy: 2

This option controls how host header and SNI name mismatches are handled. Mismatches may result in SNI-based policies defined in sni.yaml being avoided. For example, foo.com may be the fqdn value in sni.yaml which defines that client certificates are required. The user could specify bar.com as the SNI to avoid the policy requiring the client certificate but specify foo.com as the HTTP host header to still access the same object.

Therefore, if a host header would have triggered a SNI policy, it is possible that the user is trying to bypass a SNI policy if the host header and SNI values do not match.

If this setting is 0, no checking is performed. If this setting is 1 or 2, the host header and SNI values are compared if the host header value would have triggered a SNI policy. If there is a mismatch and the value is 1, a warning is generated but the transaction is allowed to proceed. If the value is 2 and there is a mismatch, a warning is generated and a status 403 is returned.

Note that SNI and hostname consistency checking is not performed on all connections indiscriminately, even if this global proxy.config.http.host_sni_policy is set to a value of 1 or 2. It is only performed for connections to hosts specifying verify_client and/or ip_allow policies in sni.yaml. That is, the SNI and hostname mismatch check is only performed if a relevant security policy for the SNI is set in sni.yaml. The proxy.config.http.host_sni_policy records.yaml value is used as the default value if either of these policies is set in the corresponding sni.yaml file entry and the sni.yaml entry does not override this value via a host_sni_policy attribute.

IP Allow

proxy.config.cache.ip_allow.filename

Scope:

CONFIG

Type:

STRING

Default:

ip_allow.yaml

Reloadable:

Yes

yaml:

records:
  cache:
    ip_allow:
      filename: ip_allow.yaml

Set the file path for the IP allow configuration file. For details of the use of this file, see ip_allow.yaml. If this is a relative path, Traffic Server loads it relative to the SYSCONFDIR directory.

proxy.config.cache.ip_categories.filename

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  cache:
    ip_categories:
      filename: null

Set the file path for the IP allow categories definition file. For details of the use of this file, see ip_allow.yaml. If this is a relative path, Traffic Server loads it relative to the SYSCONFDIR directory.

Cache Control

proxy.config.cache.enable_read_while_writer

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  cache:
    enable_read_while_writer: 1

Specifies when to enable the ability to read a cached object while another connection is completing the write to cache for that same object. The goal here is to avoid multiple origin connections for the same cacheable object upon a cache miss. The possible values of this config are:

Value	Description
0	Never read while writing.
1	Always read while writing.
2	Always read while writing, but allow non-cached Range requests through to the origin server.

The 2 option is useful to avoid delaying requests which can not easily be satisfied by the partially written response.

Several other configuration values need to be set for this to be usable. See Reducing Origin Server Requests (Avoiding the Thundering Herd).

proxy.config.cache.read_while_writer.max_retries

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

yaml:

records:
  cache:
    read_while_writer:
      max_retries: 10

Specifies how many retries trafficserver attempts to trigger read_while_writer on failing to obtain the write VC mutex or until the first fragment is downloaded for the object being downloaded. The retry duration is specified using the setting proxy.config.cache.read_while_writer_retry.delay

proxy.config.cache.read_while_writer_retry.delay

Scope:

CONFIG

Type:

INT

Default:

50

Reloadable:

Yes

yaml:

records:
  cache:
    read_while_writer_retry:
      delay: 50

Specifies the delay in msec, trafficserver waits to reattempt read_while_writer on failing to obtain the write VC mutex or until the first fragment is downloaded for the object being downloaded. Note that trafficserver implements a progressive delay in reattempting, by doubling the configured duration from the third reattempt onwards.

proxy.config.cache.force_sector_size

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  cache:
    force_sector_size: 0

Forces the use of a specific hardware sector size, e.g. 4096, for all disks.

SSDs and “advanced format” drives claim a sector size of 512; however, it is safe to force a higher size than the hardware supports natively as we count atomicity in 512 byte increments.

4096-sized drives formatted for Windows will have partitions aligned on 63 512-byte sector boundaries, so they will be unaligned. There are workarounds, but you need to do some research on your particular drive. Some drives have a one-time option to switch the partition boundary, while others might require reformatting or repartitioning.

To be safe in Linux, you could just use the entire drive: /dev/sdb instead of /dev/sdb1 and Traffic Server will do the right thing. Misaligned partitions on Linux are auto-detected.

For example: If /sys/block/sda/sda1/alignment_offset is non-zero, ATS will offset reads/writes to that disk by that alignment. If Linux knows about any existing partition misalignments, ATS will compensate.

Partitions formatted to support hardware sector size of more than 512 (e.g. 4096) will result in all objects stored in the cache to be integral multiples of 4096 bytes, which will result in some waste for small files.

proxy.config.http.cache.http

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      http: 1

Enables (1) or disables (0) caching of HTTP requests.

proxy.config.http.cache.post_method

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      post_method: 0

Enables (1) or disables (0) caching of HTTP POST requests.

proxy.config.http.cache.generation

Scope:

CONFIG

Type:

INT

Default:

-1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      generation: -1

If set to a value other than -1, the value if this configuration option is combined with the cache key at cache lookup time. Changing this value has the effect of an instantaneous, zero-cost cache purge since it will cause all subsequent cache keys to change. Since this is an overridable configuration, it can be used to purge the entire cache, or just a specific remap.config rule.

proxy.config.http.cache.ignore_query

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_query: 0

If this value is set to 1, then the query string is ignored when calculating the cach key for the request. This can be noticeably faster than using e.g. the cachekey plugin to just remove the query parameters.

proxy.config.http.doc_in_cache_skip_dns

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    doc_in_cache_skip_dns: 1

When enabled (1), do not perform origin server DNS resolution if a fresh copy of the requested document is available in the cache. This setting has no effect if HTTP caching is disabled or if there are IP based ACLs configured.

Note that plugins, particularly authorization plugins, which use the TS_HTTP_OS_DNS_HOOK hook may require this configuration variable to be disabled (0) in order to function properly. This will ensure that the hook will be evaluated and plugin execution will occur even when there is a fresh copy of the requested object in the cache (which would normally allow the DNS lookup to be skipped, thus eliminating the hook evaluation).

The downside is that the performance gain by skipping otherwise unnecessary DNS lookups is lost. Because the variable is overridable, you may retain this performance benefit for portions of your cache which do not require the use of TS_HTTP_OS_DNS_HOOK plugins, by ensuring that the setting is first disabled within only the relevant transactions. Refer to the documentation on Configuration Remap Plugin for more information.

proxy.config.http.cache.ignore_client_no_cache

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_client_no_cache: 1

When enabled (1), Traffic Server ignores client requests to bypass the cache. Specifically, Pragma: no-cache, Cache-Control: no-cache and Cache-Control: no-store in requests are ignored.

proxy.config.http.cache.ims_on_client_no_cache

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ims_on_client_no_cache: 1

When enabled (1), Traffic Server issues a conditional request to the origin server if an incoming request has a No-Cache header.

proxy.config.http.cache.ignore_server_no_cache

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_server_no_cache: 0

When enabled (1), Traffic Server ignores origin server requests to bypass the cache. Specifically, Pragma: no-cache, Cache-Control: no-cache and Cache-Control: no-store in responses are ignored.

proxy.config.http.cache.cache_responses_to_cookies

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      cache_responses_to_cookies: 1

Specifies how cookies are cached:

Value	Description
0	Do not cache any responses to cookies.
1	Cache for any content-type.
2	Cache only for image types.
3	Cache for all but text content-types.
4	Cache for all but text content-types; except origin server response without Set-Cookie or with Cache-Control: public.

proxy.config.http.cache.ignore_authentication

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_authentication: 0

When enabled (1), Traffic Server ignores WWW-Authentication headers in responses and the responses are cached.

proxy.config.http.cache.cache_urls_that_look_dynamic

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      cache_urls_that_look_dynamic: 1

Enables (1) or disables (0) caching of URLs that look dynamic, i.e.: URLs that end in .asp or contain a question mark (?), a semicolon (;), or cgi. For a full list, please refer to HttpTransact::url_looks_dynamic

proxy.config.http.cache.when_to_revalidate

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      when_to_revalidate: 0

Specifies when to revalidate content:

Value	Description
0	Use cache directives or heuristic (the default value).
1	Stale if heuristic.
2	Always stale (always revalidate).
3	Never stale.
4	Use cache directives or heuristic (0) unless the request has an If-Modified-Since header.

If the request contains the If-Modified-Since header, then Traffic Server always revalidates the cached content and uses the client’s If-Modified-Since header for the proxy request.

proxy.config.http.cache.required_headers

Scope:

CONFIG

Type:

INT

Default:

2

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      required_headers: 2

The type of headers required in a request for the request to be cacheable.

Value	Description
0	No headers required to make document cacheable.
1	Either the Last-Modified header, or an explicit lifetime header (Expires or Cache-Control: max-age) is required.
2	Explicit lifetime is required, from either Expires or Cache-Control: max-age.

proxy.config.http.cache.max_stale_age

Scope:

CONFIG

Type:

INT

Default:

604800

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      max_stale_age: 604800

The maximum age in seconds allowed for a stale response before it cannot be cached.

proxy.config.http.cache.guaranteed_min_lifetime

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      guaranteed_min_lifetime: 0

Establishes a guaranteed minimum lifetime boundary for object freshness. Setting this to 0 (default) disables the feature.

proxy.config.http.cache.guaranteed_max_lifetime

Scope:

CONFIG

Type:

INT

Default:

31536000

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      guaranteed_max_lifetime: 31536000

Establishes a guaranteed maximum lifetime boundary for object freshness. Setting this to 0 disables the feature.

proxy.config.http.cache.range.lookup

Scope:

CONFIG

Type:

INT

Default:

1

Overridable:

Yes

yaml:

records:
  http:
    cache:
      range:
        lookup: 1

When enabled (1), Traffic Server looks up range requests in the cache.

proxy.config.http.cache.range.write

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  http:
    cache:
      range:
        write: 0

When enabled (1), Traffic Server will attempt to write (lock) the URL to cache for a request specifying a range. This is useful when the origin server might ignore a range request and respond with a full (200) response. Additionally, this setting will attempt to transform a 200 response from the origin server to a partial (206) response, honoring the requested range, while caching the full response.

proxy.config.http.cache.ignore_accept_mismatch

Scope:

CONFIG

Type:

INT

Default:

2

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_accept_mismatch: 2

When enabled with a value of 1, Traffic Server serves documents from cache with a Content-Type: header even if it does not match the Accept: header of the request. If set to 2 (default), this logic only happens in the absence of a Vary header in the cached response (which is the recommended and safe use).

Note

This option should only be enabled with 1 if you’re having problems with caching and you origin server doesn’t set the Vary header. Alternatively, if the origin is incorrectly setting Vary: Accept or doesn’t respond with 406 (Not Acceptable), you can also enable this configuration with a 1.

proxy.config.http.cache.ignore_accept_language_mismatch

Scope:

CONFIG

Type:

INT

Default:

2

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_accept_language_mismatch: 2

When enabled with a value of 1, Traffic Server serves documents from cache with a Content-Language: header even if it does not match the Accept-Language: header of the request. If set to 2 (default), this logic only happens in the absence of a Vary header in the cached response (which is the recommended and safe use).

Note

This option should only be enabled with 1 if you’re having problems with caching and you origin server doesn’t set the Vary header. Alternatively, if the origin is incorrectly setting Vary: Accept-Language or doesn’t respond with 406 (Not Acceptable), you can also enable this configuration with a 1.

proxy.config.http.cache.ignore_accept_encoding_mismatch

Scope:

CONFIG

Type:

INT

Default:

2

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_accept_encoding_mismatch: 2

When enabled with a value of 1, Traffic Server serves documents from cache with a Content-Encoding: header even if it does not match the Accept-Encoding: header of the request. If set to 2 (default), this logic only happens in the absence of a Vary header in the cached response (which is the recommended and safe use).

Note

This option should only be enabled with 1 if you’re having problems with caching and you origin server doesn’t set the Vary header. Alternatively, if the origin is incorrectly setting Vary: Accept-Encoding or doesn’t respond with 406 (Not Acceptable) you can also enable this configuration with a 1.

proxy.config.http.cache.ignore_accept_charset_mismatch

Scope:

CONFIG

Type:

INT

Default:

2

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_accept_charset_mismatch: 2

When enabled with a value of 1, Traffic Server serves documents from cache with a Content-Type: header even if it does not match the Accept-Charset: header of the request. If set to 2 (default), this logic only happens in the absence of a Vary header in the cached response (which is the recommended and safe use).

Note

This option should only be enabled with 1 if you’re having problems with caching and you origin server doesn’t set the Vary header. Alternatively, if the origin is incorrectly setting Vary: Accept-Charset or doesn’t respond with 406 (Not Acceptable), you can also enable this configuration with a 1.

proxy.config.http.cache.ignore_client_cc_max_age

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      ignore_client_cc_max_age: 1

When enabled (1), Traffic Server ignores any Cache-Control: max-age headers from the client. This technically violates the HTTP RFC, but avoids a problem where a client can forcefully invalidate a cached object.

proxy.config.cache.max_doc_size

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  cache:
    max_doc_size: 0

Specifies the maximum object size that will be cached. 0 is unlimited.

proxy.config.cache.min_average_object_size

Scope:

CONFIG

Type:

INT

Default:

8000

yaml:

records:
  cache:
    min_average_object_size: 8000

Specifies the lower boundary of average object sizes in the cache and is used in determining the number of directory buckets to allocate for the in-memory cache directory.

proxy.config.cache.permit.pinning

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  cache:
    permit:
      pinning: 0

When enabled (1), Traffic Server will keep certain HTTP objects in the cache for a certain time as specified in cache.config.

proxy.config.cache.hit_evacuate_percent

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  cache:
    hit_evacuate_percent: 0

The size of the region (as a percentage of the total content storage in a cache stripe) in front of the write cursor that constitutes a recent access hit for evacuating the accessed object.

When an object is accessed it can be marked for evacuation, that is to be copied over the write cursor and thereby preserved from being overwritten. This is done if it is no more than a specific number of bytes in front of the write cursor. The number of bytes is a percentage of the total number of bytes of content storage in the cache stripe where the object is stored and that percentage is set by this variable.

By default, the feature is off (set to 0).

proxy.config.cache.hit_evacuate_size_limit

Scope:

CONFIG

Type:

INT

Default:

0

Units:

bytes

yaml:

records:
  cache:
    hit_evacuate_size_limit: 0

Limit the size of objects that are hit evacuated.

Objects larger than the limit are not hit evacuated. A value of 0 disables the limit.

proxy.config.cache.dir.sync_frequency

Scope:

CONFIG

Type:

INT

Default:

60

Units:

seconds

yaml:

records:
  cache:
    dir:
      sync_frequency: 60

How often we will sync the cache directory entries to disk. Note that this is a minimum time, and the actual sync may be delayed if the disks are larger than how fast we allow it to write to disk (see next options).

proxy.config.cache.dir.sync_max_writes

Scope:

CONFIG

Type:

INT

Default:

2097152

Units:

bytes

yaml:

records:
  cache:
    dir:
      sync_max_writes: 2097152

How much of a stripes cache directory we will write to disk in each write cycle. Together with the sync_delay, this controls how fast we can sync the entire directory structure to disk. The default is 2MB.

proxy.config.cache.dir.sync_delay

Scope:

CONFIG

Type:

INT

Default:

500

Units:

millisecond

yaml:

records:
  cache:
    dir:
      sync_delay: 500

How long to wait between each write cycle when syncing the cache directory to disk.

proxy.config.cache.limits.http.max_alts

Scope:

CONFIG

Type:

INT

Default:

5

yaml:

records:
  cache:
    limits:
      http:
        max_alts: 5

The maximum number of alternates that are allowed for any given URL. Disable by setting to 0.

proxy.config.cache.log.alternate.eviction

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  cache:
    log:
      alternate:
        eviction: 0

When enabled (1), Traffic Server will emit a Status level log entry every time an alternate for an object is evicted due to the number of its alternates exceeding the value of proxy.config.cache.limits.http.max_alts. The URI for the evicted alternate is included in the log. This logging may be useful to determine whether proxy.config.cache.limits.http.max_alts is tuned correctly for a given environment. It also provides visibility into alternate eviction for individual objects, which can be helpful for diagnosing unexpected Vary: header behavior from particular origins.

For further details concerning the caching of alternates, see Caching HTTP Alternates.

By default, alternate eviction logging is disabled (set to 0).

proxy.config.cache.target_fragment_size

Scope:

CONFIG

Type:

INT

Default:

1048576

yaml:

records:
  cache:
    target_fragment_size: 1048576

Sets the target size of a contiguous fragment of a file in the disk cache. When setting this, consider that larger numbers could waste memory on slow connections, but smaller numbers could increase (waste) seeks.

proxy.config.cache.alt_rewrite_max_size

Scope:

CONFIG

Type:

INT

Default:

4096

Reloadable:

Yes

yaml:

records:
  cache:
    alt_rewrite_max_size: 4096

Configures the size, in bytes, of an alternate that will be considered small enough to trigger a rewrite of the resident alt fragment within a write vector. For further details on cache write vectors, refer to the developer documentation for CacheVC.

proxy.config.cache.mutex_retry_delay

Scope:

CONFIG

Type:

INT

Default:

2

Units:

milliseconds

Reloadable:

Yes

yaml:

records:
  cache:
    mutex_retry_delay: 2

The retry delay for missing a lock on a mutex in the cache component. This is used generically for most locks, except those that have an explicit configuration for the retry delay. For instance, if the cache component is notifying another continuation of a cache event and fails to get the lock for that continuation, it will use this as the delay for the retry. This is also used from the asynchronous IO threads when IO finishes and the CacheVC lock or stripe lock is required.

proxy.config.cache.max_disk_errors

Scope:

CONFIG

Type:

INT

Default:

5

yaml:

records:
  cache:
    max_disk_errors: 5

Cache disks sometimes fail due to hardware problems. Traffic Server keeps count of the number of times it encounters I/O errors when accessing each cache disk. If the number of errors on a disk reaches this setting, Traffic Server removes that disk from the cache.

Note that the count of errors is kept in memory, and is reset to zero when Traffic Server is restarted. By default, Traffic Server will not remember which cache disk has been removed in this way when it restarts. If you wish to change this behavior and prevent known bad disks from re-joining the cache upon restart, change the setting proxy.config.cache.persist_bad_disks.

proxy.config.cache.persist_bad_disks

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  cache:
    persist_bad_disks: 0

When enabled (1), Traffic Server will remember which cache disks have been marked as failed through proxy.config.cache.max_disk_errors, even after a restart. A list of known bad cache disks is written to localstatedir/known_bad_disks.txt. If you replace the known bad disks, delete this file so that Traffic Server will use them in the cache again.

When disabled (0), Traffic Server will ignore known_bad_disks.txt.

This feature is disabled by default.

RAM Cache

proxy.config.cache.ram_cache.size

Scope:

CONFIG

Type:

INT

Default:

-1

yaml:

records:
  cache:
    ram_cache:
      size: -1

By default the RAM cache size is automatically determined, based on disk cache size; approximately 10 MB of RAM cache per GB of disk cache. Alternatively, it can be set to a fixed value such as 20GB (21474836480)

proxy.config.cache.ram_cache_cutoff

Scope:

CONFIG

Type:

INT

Default:

4194304

yaml:

records:
  cache:
    ram_cache_cutoff: 4194304

Objects greater than this size will not be kept in the RAM cache. This should be set high enough to keep objects accessed frequently in memory in order to improve performance. 4MB (4194304)

proxy.config.cache.ram_cache.algorithm

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  cache:
    ram_cache:
      algorithm: 1

Two distinct RAM caches are supported, the default (1) being the simpler LRU (Least Recently Used) cache. As an alternative, the CLFUS (Clocked Least Frequently Used by Size) is also available, by changing this configuration to 0.

proxy.config.cache.ram_cache.use_seen_filter

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  cache:
    ram_cache:
      use_seen_filter: 1

Enabling this option will filter inserts into the RAM cache to ensure that they have been seen at least once. For the LRU, this provides scan resistance.

As of ATS v10.0.0 and later, this setting can take values in the range 0 to 9. Values above 1 will only enable the seen filter after a certain threshold of RAM cache usage has been reached. The threshold is determined by the value of this setting, with 2 being 50% filled, 3 being 67% filled, and so on.

Note that CLFUS already requires that a document have history before it is inserted, so for CLFUS, setting this option means that a document must be seen three times before it is added to the RAM cache.

proxy.config.cache.ram_cache.compress

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  cache:
    ram_cache:
      compress: 0

The CLFUS RAM cache also supports an optional in-memory compression. This is not to be confused with Content-Encoding: gzip compression. The RAM cache compression is intended to try to save space in the RAM, and is not visible to the User-Agent (client).

Possible values are:

Value	Description
0	No compression
1	Fastlz (extremely fast, relatively low compression)
2	Libz (moderate speed, reasonable compression)
3	Liblzma (very slow, high compression)

Compression runs on task threads. To use more cores for RAM cache compression, increase proxy.config.task_threads.

Heuristic Expiration

proxy.config.http.cache.heuristic_min_lifetime

Scope:

CONFIG

Type:

INT

Default:

3600

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      heuristic_min_lifetime: 3600

The minimum amount of time, in seconds, an HTTP object without an expiration date can remain fresh in the cache before is considered to be stale.

proxy.config.http.cache.heuristic_max_lifetime

Scope:

CONFIG

Type:

INT

Default:

86400

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      heuristic_max_lifetime: 86400

The maximum amount of time, in seconds, an HTTP object without an expiration date can remain fresh in the cache before is considered to be stale.

proxy.config.http.cache.heuristic_lm_factor

Scope:

CONFIG

Type:

FLOAT

Default:

0.10

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      heuristic_lm_factor: 0.1

The aging factor for freshness computations. Traffic Server stores an object for this percentage of the time that elapsed since it last changed.

Dynamic Content & Content Negotiation

proxy.config.http.cache.open_read_retry_time

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      open_read_retry_time: 10

The number of milliseconds a cacheable request will wait before requesting the object from cache if an equivalent request is in flight.

proxy.config.http.cache.max_open_read_retries

Scope:

CONFIG

Type:

INT

Default:

-1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      max_open_read_retries: -1

The number of times to attempt fetching an object from cache if there was an equivalent request in flight.

proxy.config.http.cache.max_open_write_retries

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      max_open_write_retries: 1

The number of times to attempt a cache open write upon failure to get a write lock.

This config is ignored when proxy.config.http.cache.open_write_fail_action is set to 5 or proxy.config.http.cache.max_open_write_retry_timeout is set to gt 0.

proxy.config.http.cache.max_open_write_retry_timeout

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      max_open_write_retry_timeout: 0

A timeout for attempting a cache open write upon failure to get a write lock.

This config is ignored when proxy.config.http.cache.open_write_fail_action is set to 5.

proxy.config.http.cache.open_write_fail_action

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    cache:
      open_write_fail_action: 0

This setting indicates the action taken on failing to obtain the cache open write lock on either a cache miss or a cache hit stale. This typically happens when there is more than one request to the same cache object simultaneously. During such a scenario, all but one (which goes to the origin) request is served either a stale copy or an error depending on this setting.

Value	Description
0	Default. Disable cache and go to origin server.
1	Return a 502 error on a cache miss.
2	Serve stale if object’s age is under proxy.config.http.cache.max_stale_age. Otherwise, go to origin server.
3	Return a 502 error on a cache miss or serve stale on a cache revalidate if object’s age is under proxy.config.http.cache.max_stale_age. Otherwise, go to origin server.
4	Return a 502 error on either a cache miss or on a revalidation.
5	Retry Cache Read on a Cache Write Lock failure. This option together with proxy.config.cache.enable_read_while_writer configuration allows to collapse concurrent requests without a need for any plugin. Make sure to configure the Read While Writer feature correctly. Note that this option may result in CACHE_LOOKUP_COMPLETE HOOK being called back more than once.

Customizable User Response Pages

proxy.config.body_factory.enable_customizations

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  body_factory:
    enable_customizations: 1

Specifies whether customizable response pages are language specific or not:

Value	Description
1	Enable customizable user response pages in the default directory only.
2	Enable language-targeted user response pages.
3	Enable host-targeted user response pages.

proxy.config.body_factory.enable_logging

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  body_factory:
    enable_logging: 0

Enables (1) or disables (0) logging for customizable response pages. When enabled, Traffic Server records a message in the error log each time a customized response page is used or modified.

proxy.config.body_factory.template_sets_dir

Scope:

CONFIG

Type:

STRING

Default:

etc/trafficserver/body_factory

yaml:

records:
  body_factory:
    template_sets_dir: etc/trafficserver/body_factory

The customizable response page default directory. If this is a relative path, Traffic Server resolves it relative to the PREFIX directory.

proxy.config.body_factory.template_base

Scope:

CONFIG

Type:

STRING

Default:

“”

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  body_factory:
    template_base: '""'

A prefix for the file name to use to find an error template file. If set (not the empty string) this value and an underscore are prepended to the file name to find in the template sets directory. See HTML Messages Sent to Clients.

proxy.config.body_factory.response_max_size

Scope:

CONFIG

Type:

INT

Default:

8192

Reloadable:

Yes

yaml:

records:
  body_factory:
    response_max_size: 8192

Maximum size of the error template response page.

proxy.config.body_factory.response_suppression_mode

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  body_factory:
    response_suppression_mode: 0

Specifies when Traffic Server suppresses generated response pages:

Value	Description
0	Never suppress generated response pages.
1	Always suppress generated response pages.
2	Suppress response pages only for internal traffic.

DNS

proxy.config.dns.search_default_domains

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  dns:
    search_default_domains: 0

Traffic Server can attempt to resolve unqualified hostnames by expanding to the local domain. For example if a client makes a request to an unqualified host (e.g. host_x) and the Traffic Server local domain is y.com, then Traffic Server will expand the hostname to host_x.y.com.

Value	Description
0	Disable local domain expansion.
1	Enable local domain expansion.
2	Enable local domain expansion, but do not split local domain name.

proxy.config.dns.splitDNS.enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  dns:
    splitDNS:
      enabled: 0

Enables (1) or disables (0) DNS server selection. When enabled, Traffic Server refers to the splitdns.config file for the selection specification. Refer to Configuring DNS Server Selection.

proxy.config.dns.resolv_conf

Scope:

CONFIG

Type:

STRING

Default:

/etc/resolv.conf

yaml:

records:
  dns:
    resolv_conf: /etc/resolv.conf

Allows one to specify which resolv.conf file to use for finding resolvers. While the format of this file must be the same as the standard resolv.conf file, this option allows an administrator to manage the set of resolvers in an external configuration file, without affecting how the rest of the operating system uses DNS. Note that this setting works in conjunction with proxy.config.dns.nameservers, with its settings appended to the resolv.conf contents.

proxy.config.dns.round_robin_nameservers

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  dns:
    round_robin_nameservers: 1

Enables (1) or disables (0) DNS server round-robin.

proxy.config.dns.nameservers

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  dns:
    nameservers: null

The DNS servers. Note that this does not override proxy.config.dns.resolv_conf. That is, the contents of the file listed in proxy.config.dns.resolv_conf will be appended to the list of nameservers specified here. To prevent this, a bogus file can be listed there.

Example

IPv4 DNS server, loopback and port 9999

CONFIG proxy.config.dns.nameservers STRING 127.0.0.1:9999

Example

IPv6 DNS server, loopback and port 9999

CONFIG proxy.config.dns.nameservers STRING [::1]:9999

proxy.config.srv_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  srv_enabled: 0

Enables (1) or disables (0) the use of SRV records for origin server lookup. Traffic Server will use weights found in the SRV record as a weighted round robin in origin selection. Note that Traffic Server will lookup _$scheme._$internet_protocol.$origin_name. For instance, if the origin is set to https://my.example.com, Traffic Server would lookup _https._tcp.my.example.com. Also note that the port returned in the SRV record MUST match the port being used for the origin (e.g. if the origin scheme is http and a default port, there should be a SRV record with port 80).

proxy.config.dns.dedicated_thread

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  dns:
    dedicated_thread: 0

Create and dedicate a thread entirely for DNS processing. This is probably most useful on system which do a significant number of DNS lookups, typically forward proxies. But even on other systems, it can avoid some contention on the first worker thread (which otherwise takes on the burden of all DNS lookups).

proxy.config.dns.validate_query_name

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  dns:
    validate_query_name: 0

When enabled (1) provides additional resilience against DNS forgery (for instance in DNS Injection attacks), particularly in forward or transparent proxies, but requires that the resolver populates the queries section of the response properly.

proxy.config.dns.connection_mode

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  dns:
    connection_mode: 0

Three connection modes between Traffic Server and nameservers can be set – UDP_ONLY, TCP_RETRY, TCP_ONLY.

Value	Description
0	UDP_ONLY: Traffic Server always talks to nameservers over UDP.
1	TCP_RETRY: Traffic Server first UDP, retries with TCP if UDP response is truncated.
2	TCP_ONLY: Traffic Server always talks to nameservers over TCP.

proxy.config.dns.max_tcp_continuous_failures

Scope:

CONFIG

Type:

INT

Default:

10

yaml:

records:
  dns:
    max_tcp_continuous_failures: 10

If DNS connection mode is TCP_RETRY, set the threshold of the continuous TCP query failures count for the TCP connection, reset the TCP connection immediately if the continuous TCP query failures conut over the threshold. If the threshold is 0 (or less than 0) we close this feature.

proxy.config.dns.max_dns_in_flight

Scope:

CONFIG

Type:

INT

Default:

2048

yaml:

records:
  dns:
    max_dns_in_flight: 2048

Maximum inflight DNS queries made by Traffic Server at any given instant

proxy.config.dns.lookup_timeout

Scope:

CONFIG

Type:

INT

Default:

20

yaml:

records:
  dns:
    lookup_timeout: 20

Time to wait for a DNS response in seconds.

proxy.config.dns.retries

Scope:

CONFIG

Type:

INT

Default:

5

yaml:

records:
  dns:
    retries: 5

Maximum number of retries made by Traffic Server on a given DNS query

proxy.config.dns.local_ipv4

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  dns:
    local_ipv4: null

Local IPV4 address to bind to in order to make DNS requests

proxy.config.dns.local_ipv6

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  dns:
    local_ipv6: null

Local IPV6 address to bind to in order to make DNS requests

HostDB

proxy.config.hostdb.lookup_timeout

Scope:

CONFIG

Type:

INT

Default:

30

Units:

seconds

Reloadable:

Yes

yaml:

records:
  hostdb:
    lookup_timeout: 30

Time to wait for a DNS response in seconds.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.hostdb.serve_stale_for

Scope:

CONFIG

Type:

INT

Default:

*NONE*

Units:

seconds

Reloadable:

Yes

yaml:

records:
  hostdb:
    serve_stale_for: null

The number of seconds for which to use a stale NS record while initiating a background fetch for the new data.

If not set then stale records are not served.

proxy.config.hostdb.max_size

Scope:

CONFIG

Type:

INT

Default:

10737418240

Units:

bytes

yaml:

records:
  hostdb:
    max_size: 10737418240

The maximum amount of space (in bytes) allocated to hostdb. Setting this value to -1 will disable size limit enforcement.

proxy.config.hostdb.max_count

Scope:

CONFIG

Type:

INT

Default:

-1

yaml:

records:
  hostdb:
    max_count: -1

The maximum number of entries that can be stored in hostdb. A value of -1 disables item count limit enforcement.

Note

For values above 200000, you must increase proxy.config.hostdb.max_size by at least 44 bytes per entry.

proxy.config.hostdb.round_robin_max_count

Scope:

CONFIG

Type:

INT

Default:

16

yaml:

records:
  hostdb:
    round_robin_max_count: 16

The maximum count of DNS answers per round robin hostdb record. The default variable is 16.

proxy.config.hostdb.ttl_mode

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  hostdb:
    ttl_mode: 0

A host entry will eventually time out and be discarded. This variable controls how that time is calculated. A DNS request will return a TTL value and an internal value can be set with proxy.config.hostdb.timeout. This variable determines which value will be used.

Value	TTL
0	The TTL from the DNS response.
1	The internal timeout value.
2	The smaller of the DNS and internal TTL values. The internal timeout value becomes a maximum TTL.
3	The larger of the DNS and internal TTL values. The internal timeout value become a minimum TTL.

proxy.config.hostdb.timeout

Scope:

CONFIG

Type:

INT

Default:

86400

Units:

seconds

Reloadable:

Yes

yaml:

records:
  hostdb:
    timeout: 86400

Internal time to live value for host DB entries in seconds.

See proxy.config.hostdb.ttl_mode for when this value is used. See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.hostdb.fail.timeout

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  hostdb:
    fail:
      timeout: 0

Time to live value for “failed” hostdb lookups.

Note

HostDB considers any response that does not contain a response to the query a failure. This means “failure” responses (such as SOA) are subject to this timeout

proxy.config.hostdb.strict_round_robin

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  hostdb:
    strict_round_robin: 0

Set host resolution to use strict round robin.

When this and proxy.config.hostdb.timed_round_robin are both disabled (set to 0), Traffic Server always uses the same origin server for the same client, for as long as the origin server is available. Otherwise if this is set then IP address is rotated on every request. This setting takes precedence over proxy.config.hostdb.timed_round_robin.

proxy.config.hostdb.timed_round_robin

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  hostdb:
    timed_round_robin: 0

Set host resolution to use timed round robin.

When this and proxy.config.hostdb.strict_round_robin are both disabled (set to 0), Traffic Server always uses the same origin server for the same client, for as long as the origin server is available. Otherwise if this is set to N the IP address is rotated if more than N seconds have passed since the first time the current address was used.

proxy.config.hostdb.host_file.path

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  hostdb:
    host_file:
      path: null

Set the file path for an external host file.

If this is set (non-empty) then the file is presumed to be a hosts file in the standard . It is read and the entries there added to the HostDB. The file is periodically checked for a more recent modification date in which case it is reloaded. The interval is set with proxy.config.hostdb.host_file.interval.

While not technically reloadable, the value is read every time the file is to be checked so that if changed the new value will be used on the next check and the file will be treated as modified.

proxy.config.hostdb.host_file.interval

Scope:

CONFIG

Type:

INT

Default:

86400

Units:

seconds

Reloadable:

Yes

yaml:

records:
  hostdb:
    host_file:
      interval: 86400

Set the file changed check timer for proxy.config.hostdb.host_file.path.

The file is checked every this many seconds to see if it has changed. If so the HostDB is updated with the new values in the file.

proxy.config.hostdb.partitions

Scope:

CONFIG

Type:

INT

Default:

64

yaml:

records:
  hostdb:
    partitions: 64

The number of partitions for hostdb. If you are seeing lock contention within hostdb’s cache (due to a large number of records) you can increase the number of partitions

proxy.config.hostdb.ip_resolve

Scope:

CONFIG

Type:

STRING

Default:

NULL

Overridable:

Yes

yaml:

records:
  hostdb:
    ip_resolve: null

Set the host resolution style.

This is an ordered list of keywords separated by semicolons that specify how a host name is to be resolved to an IP address. The keywords are case insensitive.

Keyword	Description
ipv4	Resolve to an IPv4 address.
ipv6	Resolve to an IPv6 address.
client	Resolve to the same family as the client IP address.
only	Stop resolving.

The order of the keywords is critical. When a host name needs to be resolved it is resolved in same order as the keywords. If a resolution fails, the next option in the list is tried. The keyword only means to give up resolution entirely. The keyword list has a maximum length of three keywords, more are never needed. By default there is an implicit ipv4;ipv6 attached to the end of the string unless the keyword only appears.

Example

Use the incoming client family, then try IPv4 and IPv6.

client;ipv4;ipv6

Because of the implicit resolution this can also be expressed as just

client

Example

Resolve only to IPv4.

ipv4;only

Example

Resolve only to the same family as the client (do not permit cross family transactions).

client;only

This value is a global default that can be overridden by proxy.config.http.server_ports.

Note

This style is used as a convenience for the administrator. During a resolution the resolution order will be one family, then possibly the other. This is determined by changing client to ipv4 or ipv6 based on the client IP address and then removing duplicates.

Important

This option has no effect on outbound transparent connections The local IP address used in the connection to the origin server is determined by the client, which forces the IP address family of the address used for the origin server. In effect, outbound transparent connections always use a resolution style of “client”.

proxy.config.hostdb.verify_after

Scope:

CONFIG

Type:

INT

Default:

720

yaml:

records:
  hostdb:
    verify_after: 720

Set the interval (in seconds) in which to re-query DNS regardless of TTL status.

Logging Configuration

proxy.config.log.logging_enabled

Scope:

CONFIG

Type:

INT

Default:

3

Reloadable:

Yes

yaml:

records:
  log:
    logging_enabled: 3

Enables and disables event logging:

Value	Effect
0	Logging disabled.
1	Log errors only.
2	Log transactions only.
3	Dual logging (errors and transactions).

Refer to Logging for more information on event logging.

proxy.config.log.log_fast_buffer

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  log:
    log_fast_buffer: 0

Enables fast logging mode as the default for all log objects. This mode can log larger transaction rates, but log entries will appear out of order in the log output. You can enable fast mode for individual log objects in logging.yaml file by adding fast: true to that object’s config.

proxy.config.log.max_secs_per_buffer

Scope:

CONFIG

Type:

INT

Default:

5

Reloadable:

Yes

yaml:

records:
  log:
    max_secs_per_buffer: 5

The maximum amount of time before data in the buffer is flushed to disk.

Note

The effective lower bound to this config is whatever proxy.config.log.periodic_tasks_interval is set to.

proxy.config.log.max_space_mb_for_logs

Scope:

CONFIG

Type:

INT

Default:

25000

Units:

megabytes

Reloadable:

Yes

yaml:

records:
  log:
    max_space_mb_for_logs: 25000

The amount of space allocated to the logging directory (in MB). The headroom amount specified by proxy.config.log.max_space_mb_headroom is taken from this space allocation.

Note

All files in the logging directory contribute to the space used, even if they are not log files.

proxy.config.log.max_space_mb_headroom

Scope:

CONFIG

Type:

INT

Default:

1000

Units:

megabytes

Reloadable:

Yes

yaml:

records:
  log:
    max_space_mb_headroom: 1000

The tolerance for the log space limit (in megabytes). If the variable proxy.config.log.auto_delete_rolled_files is set to 1 (enabled), then auto-deletion of log files is triggered when the amount of free space available in the logging directory is less than the value specified here.

proxy.config.log.hostname

Scope:

CONFIG

Type:

STRING

Default:

localhost

Reloadable:

Yes

yaml:

records:
  log:
    hostname: localhost

The hostname of the machine running Traffic Server.

proxy.config.log.logfile_dir

Scope:

CONFIG

Type:

STRING

Default:

var/log/trafficserver

Reloadable:

Yes

yaml:

records:
  log:
    logfile_dir: var/log/trafficserver

The path to the logging directory. This can be an absolute path or a path relative to the PREFIX directory in which Traffic Server is installed.

Note

The directory you specify must already exist.

proxy.config.log.logfile_perm

Scope:

CONFIG

Type:

STRING

Default:

rw-r–r–

Reloadable:

Yes

yaml:

records:
  log:
    logfile_perm: rw-r--r--

The log file permissions. The standard UNIX file permissions are used (owner, group, other). Permissible values are:

Value	Description
-	No permissions.
r	Read permission.
w	Write permission.
x	Execute permission.

Permissions are subject to the umask settings for the Traffic Server process. This means that a umask setting of 002 will not allow write permission for others, even if specified in the configuration file. Permissions for existing log files are not changed when the configuration is modified.

proxy.config.log.rolling_enabled

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  log:
    rolling_enabled: 1

Specifies how log files are rolled. You can specify the following values:

Value	Description
0	Disables log file rolling.
1	Enables log file rolling at specific intervals during the day (specified with the proxy.config.log.rolling_interval_sec and proxy.config.log.rolling_offset_hr variables).
2	Enables log file rolling when log files reach a specific size (specified with proxy.config.log.rolling_size_mb).
3	Enables log file rolling at specific intervals during the day or when log files reach a specific size (whichever occurs first).
4	Enables log file rolling at specific intervals during the day when log files reach a specific size (i.e. at a specified time if the file is of the specified size).

proxy.config.log.rolling_interval_sec

Scope:

CONFIG

Type:

INT

Default:

86400

Reloadable:

Yes

yaml:

records:
  log:
    rolling_interval_sec: 86400

The log file rolling interval, in seconds. The minimum value is 60 (1 minute). The maximum, and default, value is 86400 seconds (one day).

Note

If you start Traffic Server within a few minutes of the next rolling time, then rolling might not occur until the next rolling time.

proxy.config.log.rolling_offset_hr

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  log:
    rolling_offset_hr: 0

The file rolling offset hour. The hour of the day that starts the log rolling period.

proxy.config.log.rolling_size_mb

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

yaml:

records:
  log:
    rolling_size_mb: 10

The size, in megabytes, that log files must reach before rolling takes place. The minimum value for this setting is 10.

proxy.config.log.rolling_min_count

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  log:
    rolling_min_count: 0

Specifies the minimum count of rolled (event) logs to keep. This value will be used to decide the order of auto-deletion (if enabled). A default value of 0 means auto-deletion will try to keep logs as much as possible. This value can be and should be overridden in logging.yaml. See Log Rotation and Retention for guidance.

proxy.config.log.rolling_max_count

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  log:
    rolling_max_count: 0

Specifies the maximum count of rolled output logs to keep. This value will be used by the auto-deletion (if enabled) to trim the number of rolled log files every time the log is rolled. A default value of 0 means auto-deletion will not try to limit the number of output logs. See Log Rotation and Retention for an use-case for this option.

proxy.config.log.rolling_allow_empty

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  log:
    rolling_allow_empty: 0

While rolling default behavior is to rename, close and re-open the log file only when/if there is something to log to the log file. This option opens a new log file right after rolling even if there is nothing to log (i.e. nothing to be logged due to lack of requests to the server) which may lead to 0-sized log files while rolling. See Log Rotation and Retention for an use-case for this option.

Value	Description
0	No empty log files created and rolled if there was nothing to log
1	Allow empty log files to be created and rolled even if there was nothing to log

proxy.config.log.auto_delete_rolled_files

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  log:
    auto_delete_rolled_files: 1

Enables (1) or disables (0) automatic deletion of rolled files.

proxy.config.log.sampling_frequency

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  log:
    sampling_frequency: 1

Configures Traffic Server to log only a sample of transactions rather than every transaction. You can specify the following values:

Value	Description
1	Log every transaction.
2	Log every second transaction.
3	Log every third transaction.
n	… and so on…

proxy.config.log.periodic_tasks_interval

Scope:

CONFIG

Type:

INT

Default:

5

Units:

seconds

Reloadable:

Yes

yaml:

records:
  log:
    periodic_tasks_interval: 5

How often Traffic Server executes log related periodic tasks, in seconds

proxy.config.log.proxy.config.log.throttling_interval_msec

Scope:

CONFIG

Type:

INT

Default:

60000

Units:

milliseconds

Reloadable:

Yes

yaml:

records:
  log:
    proxy:
      config:
        log:
          throttling_interval_msec: 60000

The minimum amount of milliseconds between repeated throttled Traffic Server log events. A value of 0 implies no throttling. Note that for performance reasons only certain logs are compiled with throttling applied to them.

Throttling is applied to all log events for a particular message which is emitted within its throttling interval. That is, once a throttled log is emitted, none will be emitted until the next log event for that message which occurs outside of this configured interval. As mentioned above, this message is applied not broadly but rather to potentially noisy log messages, such as ones that might occur thousands of times a second under certain error conditions. Once the next log event occurs outside of its interval, a summary message is printed conveying how many messages of that type were throttled since the last time it was emitted.

It is possible that a log is emitted, followed by more of its type in an interval, then none are emitted after that. Be aware this would result in no summary log message for that interval until the message is emitted again outside of the throttled interval.

proxy.config.http.slow.log.threshold

Scope:

CONFIG

Type:

INT

Default:

0

Units:

milliseconds

Reloadable:

Yes

yaml:

records:
  http:
    slow:
      log:
        threshold: 0

If set to a non-zero value N then any connection that takes longer than N milliseconds from accept to completion will cause its timing stats to be written to the debugging log file. This is identifying data about the transaction and all of the transaction milestones.

proxy.config.http2.connection.slow.log.threshold

Scope:

CONFIG

Type:

INT

Default:

0

Units:

milliseconds

Reloadable:

Yes

yaml:

records:
  http2:
    connection:
      slow:
        log:
          threshold: 0

If set to a non-zero value N then any HTTP/2 connection that takes longer than N milliseconds from open to close will cause its timing stats to be written to the debugging log file. This is identifying data about the transaction and all of the transaction milestones.

proxy.config.http2.stream.slow.log.threshold

Scope:

CONFIG

Type:

INT

Default:

0

Units:

milliseconds

Reloadable:

Yes

yaml:

records:
  http2:
    stream:
      slow:
        log:
          threshold: 0

If set to a non-zero value N then any HTTP/2 stream that takes longer than N milliseconds from open to close will cause its timing stats to be written to the debugging log file. This is identifying data about the transaction and all of the transaction milestones.

proxy.config.log.config.filename

Scope:

CONFIG

Type:

STRING

Default:

logging.yaml

Reloadable:

Yes

Deprecated:

Yes

yaml:

records:
  log:
    config:
      filename: logging.yaml

This configuration value specifies the path to the logging.yaml configuration file. If this is a relative path, Traffic Server loads it relative to the SYSCONFDIR directory.

proxy.config.log.max_line_size

Scope:

CONFIG

Type:

INT

Default:

9216

Units:

bytes

yaml:

records:
  log:
    max_line_size: 9216

This controls the maximum line length for ASCII formatted log entries. This applies to ASCII_PIPE and ASCII file logs, unless proxy.config.log.ascii_buffer_size is also specified and the value of ascii_buffer_size is larger than max_line_size: in that case, max_line_size only applies to ASCII_PIPE logs while ascii_buffer_size will apply to ASCII (non-pipe) log files.

proxy.config.log.ascii_buffer_size

Scope:

CONFIG

Type:

INT

Default:

36864

Units:

bytes

yaml:

records:
  log:
    ascii_buffer_size: 36864

This controls the maximum line length for ASCII formatted log entries that are non-pipe log files. If this value is smaller than proxy.config.log.max_line_size, then the latter will be used for both ASCII and ASCII_PIPE log files. If both max_line_size and ascii_buffer_size are set, then max_line_size will be used for ASCII_PIPE logs while ascii_buffer_size will be used for ASCII (non-pipe) log files. This all might seem complicated, but just keep in mind that the intention of ascii_buffer_size is to simply provide a way for the user to configure different ASCII and ASCII_PIPE maximum line lengths.

proxy.config.log.log_buffer_size

Scope:

CONFIG

Type:

INT

Default:

9216

Units:

bytes

Reloadable:

Yes

yaml:

records:
  log:
    log_buffer_size: 9216

This is an orthogonal mechanism from proxy.config.log.max_line_size and proxy.config.log.ascii_buffer_size for limiting line length size by constraining the log entry buffer to a particular amount of memory. Unlike the above two configurations, log_buffer_size applies to both binary and ASCII log file entries. For ASCII log files, if a maximum log size is set via both the above mechanisms and by log_buffer_size, then the smaller of the two configurations will be applied to the line length.

Diagnostic Logging Configuration

proxy.config.diags.output.diag

Scope:

CONFIG

Type:

STRING

Default:

E

yaml:

records:
  diags:
    output:
      diag: E

proxy.config.diags.output.debug

Scope:

CONFIG

Type:

STRING

Default:

E

yaml:

records:
  diags:
    output:
      debug: E

proxy.config.diags.output.status

Scope:

CONFIG

Type:

STRING

Default:

L

yaml:

records:
  diags:
    output:
      status: L

proxy.config.diags.output.note

Scope:

CONFIG

Type:

STRING

Default:

L

yaml:

records:
  diags:
    output:
      note: L

proxy.config.diags.output.warning

Scope:

CONFIG

Type:

STRING

Default:

L

yaml:

records:
  diags:
    output:
      warning: L

proxy.config.diags.output.error

Scope:

CONFIG

Type:

STRING

Default:

SL

yaml:

records:
  diags:
    output:
      error: SL

proxy.config.diags.output.fatal

Scope:

CONFIG

Type:

STRING

Default:

SL

yaml:

records:
  diags:
    output:
      fatal: SL

proxy.config.diags.output.alert

Scope:

CONFIG

Type:

STRING

Default:

L

yaml:

records:
  diags:
    output:
      alert: L

proxy.config.diags.output.emergency

Scope:

CONFIG

Type:

STRING

Default:

SL

yaml:

records:
  diags:
    output:
      emergency: SL

The diagnostic output configuration variables control where Traffic Server should log diagnostic output. Messages at each diagnostic level can be directed to any combination of diagnostic destinations. Valid diagnostic message destinations are:

Value	Description
O	Log to standard output.
E	Log to standard error.
S	Log to syslog.
L	Log to diags.log (with the filename configurable via proxy.config.diags.logfile.filename).

Example

To log debug diagnostics to both syslog and diags.log:

CONFIG proxy.config.diags.output.debug STRING SL

proxy.config.diags.show_location

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  diags:
    show_location: 1

Annotates diagnostic messages with the source code location. Set to 1 to enable for Dbg() messages only. Set to 2 to enable for all messages.

proxy.config.diags.debug.enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  diags:
    debug:
      enabled: 0

When set to 1, enables logging for diagnostic messages whose log level is diag or debug.

When set to 2, interprets the proxy.config.diags.debug.client_ip setting determine whether diagnostic messages are logged.

See the Enable debug using traffic_ctl for a convenient way to handle this.

proxy.config.diags.debug.client_ip

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  diags:
    debug:
      client_ip: null

if proxy.config.diags.debug.enabled is set to 2, this value is tested against the source IP of the incoming connection. If there is a match, all the diagnostic messages for that connection and the related outgoing connection will be logged.

proxy.config.diags.debug.tags

Scope:

CONFIG

Type:

STRING

Default:

http|dns

yaml:

records:
  diags:
    debug:
      tags: http|dns

Each Traffic Server diag and debug level message is annotated with a subsystem tag. This configuration contains an anchored regular expression that filters the messages based on the tag. The expressions are prefix matched which creates an implicit .* at the end. Therefore the default value http|dns will match tags such as http, http_hdrs, dns, and dns_recv.

Some commonly used debug tags are:

Tag	Subsystem usage
dns	DNS query resolution
http_hdrs	Logs the headers for HTTP requests and responses
privileges	Privilege elevation
ssl	TLS termination and certificate processing

proxy.config.diags.debug.throttling_interval_msec

Scope:

CONFIG

Type:

INT

Default:

0

Units:

milliseconds

Reloadable:

Yes

yaml:

records:
  diags:
    debug:
      throttling_interval_msec: 0

The minimum amount of milliseconds between repeated Traffic Server diag and debug log events. A value of 0 implies no throttling. All diags and debug logs are compiled with throttling applied to them.

For details about how log throttling works, see log.throttling_interval_msec.

proxy.config.diags.logfile.filename

Scope:

CONFIG

Type:

STRING

Default:

diags.log

yaml:

records:
  diags:
    logfile:
      filename: diags.log

The name of the file to which Traffic Server diagnostic logs will be emitted. For information on the diagnostic log file, see diags.log. For the configurable parameters concerning what log content is emitted to diags.log, see the Diagnostic Output Configuration Variables above.

If this is set to stdout or stderr, then all diagnostic logging will go to the stdout or stderr stream, respectively.

proxy.config.error.logfile.filename

Scope:

CONFIG

Type:

STRING

Default:

error.log

yaml:

records:
  error:
    logfile:
      filename: error.log

The name of the file to which Traffic Server transaction error logs will be emitted. For more information on these log messages, see error.log.

If this is set to stdout or stderr, then all transaction error logging will go to the stdout or stderr stream, respectively.

proxy.config.diags.logfile_perm

Scope:

CONFIG

Type:

STRING

Default:

rw-r–r–

yaml:

records:
  diags:
    logfile_perm: rw-r--r--

The log file permissions. The standard UNIX file permissions are used (owner, group, other). Permissible values are:

Value	Description
-	No permissions.
r	Read permission.
w	Write permission.
x	Execute permission.

Permissions are subject to the umask settings for the Traffic Server process. This means that a umask setting of 002 will not allow write permission for others, even if specified in the configuration file. Permissions for existing log files are not changed when the configuration is modified.

proxy.config.diags.logfile.rolling_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  diags:
    logfile:
      rolling_enabled: 0

Specifies how the diagnostics log is rolled. You can specify the following values:

Value	Description
0	Disables diagnostics log rolling.
1	Enables diagnostics log rolling at specific intervals (specified with proxy.config.diags.logfile.rolling_interval_sec). The “clock” starts ticking on Traffic Server startup.
2	Enables diagnostics log rolling when the diagnostics log reaches a specific size (specified with proxy.config.diags.logfile.rolling_size_mb).
3	Enables diagnostics log rolling at specific intervals or when the diagnostics log reaches a specific size (whichever occurs first).

proxy.config.diags.logfile.rolling_interval_sec

Scope:

CONFIG

Type:

INT

Default:

3600

Units:

seconds

Reloadable:

Yes

yaml:

records:
  diags:
    logfile:
      rolling_interval_sec: 3600

Specifies how often the diagnostics log is rolled, in seconds. The timer starts on Traffic Server startup.

proxy.config.diags.logfile.rolling_size_mb

Scope:

CONFIG

Type:

INT

Default:

100

Units:

megabytes

Reloadable:

Yes

yaml:

records:
  diags:
    logfile:
      rolling_size_mb: 100

Specifies at what size to roll the diagnostics log at.

proxy.config.diags.logfile.rolling_min_count

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  diags:
    logfile:
      rolling_min_count: 0

Specifies the minimum count of rolled diagnostic logs to keep. This value will be used to decide the order of auto-deletion (if enabled). A default value of 0 means auto-deletion will try to keep diagnostic logs as much as possible. See Log Rotation and Retention for guidance.

Reverse Proxy

proxy.config.reverse_proxy.enabled

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  reverse_proxy:
    enabled: 1

Enables (1) or disables (0) HTTP reverse proxy.

proxy.config.header.parse.no_host_url_redirect

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  header:
    parse:
      no_host_url_redirect: null

The URL to which to redirect requests with no host headers (reverse proxy).

URL Remap Rules

proxy.config.url_remap.filename

Scope:

CONFIG

Type:

STRING

Default:

remap.config

Deprecated:

Yes

yaml:

records:
  url_remap:
    filename: remap.config

Sets the name of the remap.config file.

proxy.config.url_remap.remap_required

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

yaml:

records:
  url_remap:
    remap_required: 1

Set this variable to 1 if you want Traffic Server to serve requests only from origin servers listed in the mapping rules of the remap.config file. If a request does not match, then the browser will receive an error.

proxy.config.url_remap.pristine_host_hdr

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  url_remap:
    pristine_host_hdr: 0

Set this variable to 1 if you want to retain the client host header in a request during remapping.

proxy.config.url_remap.min_rules_required

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  url_remap:
    min_rules_required: 0

The minimum number of rules remap.config must have to be considered valid. An otherwise valid configuration with fewer than this many rules is considered to be invalid as if it had a syntax error. A value of zero allows remap.config to be empty or absent.

This is dynamic to enable different requirements for startup and reloading.

proxy.config.url_remap.acl_behavior_policy

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  url_remap:
    acl_behavior_policy: 0

This controls how the ACL filter allow and deny actions behave in remap.config. See ACL Filters for more details.

Value	Description
0	Legacy (9.x and below) behavior.
1	Modern (10.x and above) behavior.

The value 0 provides ACL filter allow and deny action behavior that is backwards compatible with previous versions of Traffic Server. The value 1 results in a fatal log message if allow or deny is used with a message encouraging the user to transition to either set_allow or set_deny or add_allow or add_deny actions. 0 is the default value.

Note

This configuration is deprecated in 10.x. Starting with 11.x, Traffic Server will always function like this configuration is set to 1 (modern) and the configuration will be removed entirely.

SSL Termination

proxy.config.ssl.server.cipher_suite

Scope:

CONFIG

Type:

STRING

Default:

<see notes>

yaml:

records:
  ssl:
    server:
      cipher_suite: <see notes>

Configures the set of encryption, digest, authentication, and key exchange algorithms provided by OpenSSL which Traffic Server will use for SSL connections. For the list of algorithms and instructions on constructing an appropriately formatting cipher_suite string, see OpenSSL Ciphers.

The current default is:

ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-CCM:ECDHE-ECDSA-AES128-CCM:ECDHE-ECDSA-AES256-CCM8:ECDHE-ECDSA-AES128-CCM8:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES256-SHA384:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-RSA-AES256-SHA384:ECDHE-RSA-AES128-SHA256:DHE-RSA-AES256-GCM-SHA384:DHE-RSA-AES128-GCM-SHA256:DHE-RSA-AES256-CCM8:DHE-RSA-AES128-CCM8:DHE-RSA-AES256-CCM:DHE-RSA-AES128-CCM:DHE-RSA-AES256-SHA256:DHE-RSA-AES128-SHA256:AES256-GCM-SHA384:AES128-GCM-SHA256:AES256-CCM8:AES128-CCM8:AES256-CCM:AES128-CCM:AES256-SHA256:AES128-SHA2

proxy.config.ssl.client.cipher_suite

Scope:

CONFIG

Type:

STRING

Default:

<See notes under proxy.config.ssl.server.cipher_suite.>

yaml:

records:
  ssl:
    client:
      cipher_suite: <See notes under proxy.config.ssl.server.cipher_suite.>

Configures the cipher_suite which Traffic Server will use for SSL connections to origin or next hop. This currently defaults to:

ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384:DHE-DSS-AES256-GCM-SHA384:DHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-CCM8:ECDHE-ECDSA-AES256-CCM:DHE-RSA-AES256-CCM8:DHE-RSA-AES256-CCM:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:DHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-ARIA256-GCM-SHA384:ECDHE-ARIA256-GCM-SHA384:DHE-DSS-ARIA256-GCM-SHA384:DHE-RSA-ARIA256-GCM-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA384:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA256:ECDHE-ECDSA-CAMELLIA256-SHA384:ECDHE-RSA-CAMELLIA256-SHA384:DHE-RSA-CAMELLIA256-SHA256:DHE-DSS-CAMELLIA256-SHA256:RSA-PSK-AES256-GCM-SHA384:RSA-PSK-CHACHA20-POLY1305:RSA-PSK-ARIA256-GCM-SHA384:AES256-GCM-SHA384:AES256-CCM8:AES256-CCM:ARIA256-GCM-SHA384:AES256-SHA256:CAMELLIA256-SHA256:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:DHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-CCM8:ECDHE-ECDSA-AES128-CCM:DHE-RSA-AES128-CCM8:DHE-RSA-AES128-CCM:ECDHE-ECDSA-ARIA128-GCM-SHA256:ECDHE-ARIA128-GCM-SHA256:DHE-DSS-ARIA128-GCM-SHA256:DHE-RSA-ARIA128-GCM-SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA256:DHE-DSS-AES128-SHA256:ECDHE-ECDSA-CAMELLIA128-SHA256:ECDHE-RSA-CAMELLIA128-SHA256:DHE-RSA-CAMELLIA128-SHA256:DHE-DSS-CAMELLIA128-SHA256:RSA-PSK-AES128-GCM-SHA256:RSA-PSK-ARIA128-GCM-SHA256:AES128-GCM-SHA256:AES128-CCM8:AES128-CCM:ARIA128-GCM-SHA256:AES128-SHA256:CAMELLIA128-SHA256

proxy.config.ssl.server.TLSv1_3.cipher_suites

Scope:

CONFIG

Type:

STRING

Default:

<See notes>

yaml:

records:
  ssl:
    server:
      TLSv1_3:
        cipher_suites: <See notes>

Configures the pair of the AEAD algorithm and hash algorithm to be used with HKDF provided by OpenSSL which Traffic Server will use for TLSv1.3 connections. For the list of algorithms and instructions, see The -ciphersuites section of OpenSSL Ciphers.

The current default value is:

TLS_AES_256_GCM_SHA384:TLS_AES_128_GCM_SHA256:TLS_CHACHA20_POLY1305_SHA256

This configuration works with OpenSSL v1.1.1 and above.

proxy.config.ssl.server.honor_cipher_order

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  ssl:
    server:
      honor_cipher_order: 1

By default (1) Traffic Server will use the server’s cipher suites preferences instead of the client preferences. By disabling it (0) Traffic Server will use client’s cipher suites preferences.

proxy.config.ssl.server.prioritize_chacha

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    server:
      prioritize_chacha: 0

By enabling it (1) Traffic Server will temporarily reprioritize ChaCha20-Poly1305 ciphers to the top of the server cipher list if a ChaCha20-Poly1305 cipher is at the top of the client cipher list.

This configuration works with OpenSSL v1.1.1 and above.

proxy.config.ssl.client.TLSv1_3.cipher_suites

Scope:

CONFIG

Type:

STRING

Default:

<See notes under proxy.config.ssl.server.tls.cipher_suites>

yaml:

records:
  ssl:
    client:
      TLSv1_3:
        cipher_suites: <See notes under proxy.config.ssl.server.tls.cipher_suites>

Configures the cipher_suites which Traffic Server will use for TLSv1.3 connections to origin or next hop. This configuration works with OpenSSL v1.1.1 and above.

The current default is:

TLS_AES_256_GCM_SHA384:TLS_AES_128_GCM_SHA256:TLS_CHACHA20_POLY1305_SHA256

proxy.config.ssl.server.groups_list

Scope:

CONFIG

Type:

STRING

Default:

<See notes>

yaml:

records:
  ssl:
    server:
      groups_list: <See notes>

Configures the list of supported groups provided by OpenSSL which Traffic Server will be used to determine the set of shared groups. The value is a colon separated list of group NIDs or names, for example “P-521:P-384:P-256”. For instructions, see “Groups” section of TLS1.3 - OpenSSLWiki.

The current default value with OpenSSL is:

X25519:P-256:X448:P-521:P-384

This configuration works with OpenSSL v1.0.2 and above.

proxy.config.ssl.client.groups_list

Scope:

CONFIG

Type:

STRING

Default:

<See notes under proxy.config.ssl.server.groups_list.>

yaml:

records:
  ssl:
    client:
      groups_list: <See notes under proxy.config.ssl.server.groups_list.>

Configures the list of supported groups provided by OpenSSL which Traffic Server will use for the “key_share” and “supported groups” extension of TLSv1.3 connections. The value is a colon separated list of group NIDs or names, for example “P-521:P-384:P-256”. For instructions, see “Groups” section of TLS1.3 - OpenSSLWiki.

This configuration works with OpenSSL v1.0.2 and above.

proxy.config.ssl.server.version.min

Scope:

CONFIG

Type:

INT

Default:

-1

yaml:

records:
  ssl:
    server:
      version:
        min: -1

Specifies the minimum TLS version that will be offered to clients during the TLS negotiation.

Value	Version
0	TLS 1.0
1	TLS 1.1
2	TLS 1.2
3	TLS 1.3

proxy.config.ssl.server.version.max

Scope:

CONFIG

Type:

INT

Default:

-1

yaml:

records:
  ssl:
    server:
      version:
        max: -1

Specifies the maximum TLS version that will be offered to clients during the TLS negotiation.

Value	Version
0	TLS 1.0
1	TLS 1.1
2	TLS 1.2
3	TLS 1.3

proxy.config.ssl.TLSv1

Scope:

CONFIG

Type:

INT

Default:

0

Deprecated:

Yes

yaml:

records:
  ssl:
    TLSv1: 0

This setting is deprecated in favor of proxy.config.ssl.server.version.min and proxy.config.ssl.server.version.min, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLSv1.0. If not specified, disabled by default.

proxy.config.ssl.TLSv1_1

Scope:

CONFIG

Type:

INT

Default:

0

Deprecated:

Yes

yaml:

records:
  ssl:
    TLSv1_1: 0

This setting is deprecated in favor of proxy.config.ssl.server.version.min and proxy.config.ssl.server.version.min, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLS v1.1. If not specified, disabled by default. [Requires OpenSSL v1.0.1 and higher]

Note

In order to enable TLS v1 or v1.1, additional ciphers must be added to proxy.config.ssl.client.cipher_suite. For example this list would restore the SHA1 (insecure!) cipher suites suitable for these deprecated TLS versions:

ECDHE-ECDSA-AES256-SHA:ECDHE-RSA-AES256-SHA:DHE-RSA-AES256-SHA:ECDHE-ECDSA-AES128-SHA:ECDHE-RSA-AES128-SHA:DHE-RSA-AES128-SHA:AES256-SHA:AES128-SHA

proxy.config.ssl.TLSv1_2

Scope:

CONFIG

Type:

INT

Default:

1

Deprecated:

Yes

yaml:

records:
  ssl:
    TLSv1_2: 1

This setting is deprecated in favor of proxy.config.ssl.server.version.min and proxy.config.ssl.server.version.min, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLS v1.2. If not specified, enabled by default. [Requires OpenSSL v1.0.1 and higher]

proxy.config.ssl.TLSv1_3.enabled

Scope:

CONFIG

Type:

INT

Default:

1

Deprecated:

Yes

yaml:

records:
  ssl:
    TLSv1_3:
      enabled: 1

This setting is deprecated in favor of proxy.config.ssl.server.version.min and proxy.config.ssl.server.version.min, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLS v1.3. If not specified, enabled by default. [Requires OpenSSL v1.1.1 and higher]

proxy.config.ssl.client.certification_level

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    client:
      certification_level: 0

Sets the client certification level:

Value	Description
0	Client certificates are ignored. Traffic Server does not verify client certificates during the SSL handshake. Access to Traffic Server depends on Traffic Server configuration options (such as access control lists).
1	Client certificates are optional. If a client has a certificate, then the certificate is validated. If the client does not have a certificate, then the client is still allowed access to Traffic Server unless access is denied through other Traffic Server configuration options.
2	Client certificates are required. The client must be authenticated during the SSL handshake. Clients without a certificate are not allowed to access Traffic Server.

proxy.config.ssl.server.multicert.filename

Scope:

CONFIG

Type:

STRING

Default:

ssl_multicert.config

Deprecated:

Yes

yaml:

records:
  ssl:
    server:
      multicert:
        filename: ssl_multicert.config

The location of the ssl_multicert.config file, relative to the Traffic Server configuration directory. In the following example, if the Traffic Server configuration directory is /etc/trafficserver, the Traffic Server SSL configuration file and the corresponding certificates are located in /etc/trafficserver/ssl:

CONFIG proxy.config.ssl.server.multicert.filename STRING ssl/ssl_multicert.config
CONFIG proxy.config.ssl.server.cert.path STRING etc/trafficserver/ssl
CONFIG proxy.config.ssl.server.private_key.path STRING etc/trafficserver/ssl

proxy.config.ssl.server.multicert.exit_on_load_fail

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  ssl:
    server:
      multicert:
        exit_on_load_fail: 1

By default (1), Traffic Server will not start unless all the SSL certificates listed in the ssl_multicert.config file successfully load. If false (0), SSL certificate load failures will not prevent Traffic Server from starting.

proxy.config.ssl.server.cert.path

Scope:

CONFIG

Type:

STRING

Default:

/config

yaml:

records:
  ssl:
    server:
      cert:
        path: /config

The location of the SSL certificates and chains used for accepting and validation new SSL sessions. If this is a relative path, it is appended to the Traffic Server installation PREFIX. All certificates and certificate chains listed in ssl_multicert.config will be loaded relative to this path.

proxy.config.ssl.server.private_key.path

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    server:
      private_key:
        path: null

The location of the SSL certificate private keys. Change this variable only if the private key is not located in the SSL certificate file. All private keys listed in ssl_multicert.config will be loaded relative to this path.

proxy.config.ssl.server.cert_chain.filename

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    server:
      cert_chain:
        filename: null

The name of a file containing a global certificate chain that should be used with every server certificate. This file is only used if there are certificates defined in ssl_multicert.config. Unless this is an absolute path, it is loaded relative to the path specified by proxy.config.ssl.server.cert.path.

proxy.config.ssl.server.dhparams_file

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    server:
      dhparams_file: null

The name of a file containing a set of Diffie-Hellman key exchange parameters. If not specified, 2048-bit DH parameters from RFC 5114 are used. These parameters are only used if a DHE (or EDH) cipher suite has been selected.

proxy.config.ssl.CA.cert.path

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    CA:
      cert:
        path: null

The location of the certificate authority file that client certificates will be verified against.

proxy.config.ssl.CA.cert.filename

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    CA:
      cert:
        filename: null

The filename of the certificate authority that client certificates will be verified against.

proxy.config.ssl.server.ticket_key.filename

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    server:
      ticket_key:
        filename: null

The filename of the default and global ticket key for SSL sessions. The location is relative to the proxy.config.ssl.server.cert.path directory. One way to generate this would be to run head -c48 /dev/urandom | openssl enc -base64 | head -c48 > file.ticket. Also note that OpenSSL session tickets are sensitive to the version of the ca-certificates. Once the file is changed with new tickets, use traffic_ctl config reload to begin using them.

proxy.config.ssl.servername.filename

Scope:

CONFIG

Type:

STRING

Default:

sni.yaml

Deprecated:

Yes

yaml:

records:
  ssl:
    servername:
      filename: sni.yaml

The filename of the sni.yaml configuration file. If relative, it is relative to the configuration directory.

proxy.config.ssl.max_record_size

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    max_record_size: 0

This configuration specifies the maximum number of bytes to write into a SSL record when replying over a SSL session. In some circumstances this setting can improve response latency by reducing buffering at the SSL layer. This setting can have a value between 0 and 16383 (max TLS record size).

The default of 0 means to always write all available data into a single SSL record.

A value of -1 means TLS record size is dynamically determined. The strategy employed is to use small TLS records that fit into a single TCP segment for the first ~1 MB of data, but, increase the record size to 16 KB after that to optimize throughput. The record size is reset back to a single segment after ~1 second of inactivity and the record size ramping mechanism is repeated again.

proxy.config.ssl.origin_session_cache.enabled

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  ssl:
    origin_session_cache:
      enabled: 1

This configuration enables the SSL session cache for the origin server when set to 1.

Setting to 0 disables SSL session cache for the origin server.

proxy.config.ssl.origin_session_cache.size

Scope:

CONFIG

Type:

INT

Default:

10240

yaml:

records:
  ssl:
    origin_session_cache:
      size: 10240

This configuration specifies the maximum number of entries the SSL session cache for the origin server may contain.

Setting a value less than or equal to 0 effectively disables SSL session cache for the origin server.

proxy.config.ssl.session_cache.enabled

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  ssl:
    session_cache:
      enabled: 2

Enables the SSL session cache:

Value	Description
0	Disables the session cache entirely.
1	Enables the session cache using OpenSSL’s implementation.
2	Default. Enables the session cache using Traffic Server’s implementation. This implementation should perform much better than the OpenSSL implementation.

proxy.config.ssl.session_cache.timeout

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    session_cache:
      timeout: 0

This configuration specifies the lifetime of SSL session cache entries in seconds. If it is 0, then the SSL library will use a default value, typically 300 seconds. Note: This option has no affect when using the Traffic Server session cache (option 2 in proxy.config.ssl.session_cache.enabled)

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.ssl.session_cache.auto_clear

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  ssl:
    session_cache:
      auto_clear: 1

This will set the OpenSSL auto clear flag. Auto clear is enabled by default with 1 it can be disabled by changing this setting to 0.

proxy.config.ssl.session_cache.size

Scope:

CONFIG

Type:

INT

Default:

102400

yaml:

records:
  ssl:
    session_cache:
      size: 102400

This configuration specifies the maximum number of entries the SSL session cache may contain.

proxy.config.ssl.session_cache.num_buckets

Scope:

CONFIG

Type:

INT

Default:

256

yaml:

records:
  ssl:
    session_cache:
      num_buckets: 256

This configuration specifies the number of buckets to use with the Traffic Server SSL session cache implementation. The TS implementation is a fixed size hash map where each bucket is protected by a mutex.

proxy.config.ssl.session_cache.skip_cache_on_bucket_contention

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    session_cache:
      skip_cache_on_bucket_contention: 0

This configuration specifies the behavior of the Traffic Server SSL session cache implementation during lock contention on each bucket:

Value	Description
0	Default. Don’t skip session caching when bucket lock is contented.
1	Disable the SSL session cache for a connection during lock contention.

proxy.config.ssl.server.session_ticket.enable

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  ssl:
    server:
      session_ticket:
        enable: 1

Set to 1 to enable Traffic Server to process TLS tickets for TLS session resumption.

proxy.config.ssl.server.session_ticket.number

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  ssl:
    server:
      session_ticket:
        number: 2

This configuration control the number of TLSv1.3 session tickets that are issued. Take into account that setting the value to 0 will disable session caching for TLSv1.3 connections.

Lowering this setting to 1 can be interesting when proxy.config.ssl.session_cache.enabled is enabled because otherwise for every new TLSv1.3 connection two session IDs will be inserted in the session cache. On the other hand, if proxy.config.ssl.session_cache.enabled is disabled, using the default value is recommended. In those scenarios, increasing the number of tickets could be potentially beneficial for clients performing multiple requests over concurrent TLS connections as per RFC 8446 clients SHOULDN’T reuse TLS Tickets.

For more information see https://www.openssl.org/docs/man1.1.1/man3/SSL_CTX_set_num_tickets.html

proxy.config.ssl.hsts_max_age

Scope:

CONFIG

Type:

INT

Default:

-1

Overridable:

Yes

yaml:

records:
  ssl:
    hsts_max_age: -1

This configuration specifies the max-age value that will be used when adding the Strict-Transport-Security header. The value is in seconds. A value of 0 will set the max-age value to 0 and should remove the HSTS entry from the client. A value of -1 will disable this feature and not set the header. This option is only used for HTTPS requests and the header will not be set on HTTP requests.

proxy.config.ssl.hsts_include_subdomains

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  ssl:
    hsts_include_subdomains: 0

Enables (1) or disables (0) adding the includeSubdomain value to the Strict-Transport-Security header. proxy.config.ssl.hsts_max_age needs to be set to a non -1 value for this configuration to take effect.

proxy.config.ssl.allow_client_renegotiation

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    allow_client_renegotiation: 0

This configuration specifies whether the client is able to initiate renegotiation of the SSL connection. The default of 0, means the client can’t initiate renegotiation.

proxy.config.ssl.cert.load_elevated

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    cert:
      load_elevated: 0

Enables (1) or disables (0) elevation of traffic_server privileges during loading of SSL certificates. By enabling this, SSL certificate files’ access rights can be restricted to help reduce the vulnerability of certificates.

This feature requires Traffic Server to be built with POSIX capabilities enabled.

proxy.config.ssl.handshake_timeout_in

Scope:

CONFIG

Type:

INT

Default:

30

yaml:

records:
  ssl:
    handshake_timeout_in: 30

When enabled this limits the total duration for the incoming side SSL handshake.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.ssl.keylog_file

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  ssl:
    keylog_file: null

If configured, TLS session keys for TLS connections will be logged to the specified file. This file is formatted in such a way that it can be conveniently imported into tools such as Wireshark to decrypt packet captures. This should only be used for debugging purposes since the data in the keylog file can be used to decrypt the otherwise encrypted traffic. A NULL value for this disables the feature.

This feature is disabled by default.

proxy.config.ssl.ktls.enabled

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    ktls:
      enabled: 0

Enables the use of Kernel TLS. This configuration requires OpenSSL v3.0 and above, and it must have been compiled with support for Kernel TLS.

Value	Description
0	Disables the use of Kernel TLS.
1	Enables the use of Kernel TLS..

Client-Related Configuration

proxy.config.ssl.client.verify.server.policy

Scope:

CONFIG

Type:

STRING

Default:

ENFORCED

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      verify:
        server:
          policy: ENFORCED

Configures Traffic Server to verify the origin server certificate with the Certificate Authority (CA). This configuration takes a value of DISABLED, PERMISSIVE, or ENFORCED

You can override this global setting on a per domain basis in the sni.yaml file using the verify_server_policy attribute.

You can also override via the conf_remap plugin. Those changes will take precedence over the changes in sni.yaml.

DISABLED

Server Certificate will not be verified

PERMISSIVE

The provided certificate will be verified and the connection will be established irrespective of the verification result. If verification fails the name of the server will be logged.

ENFORCED

Certificate will be verified and the connection will not be established if verification fails.

proxy.config.ssl.client.verify.server.properties

Scope:

CONFIG

Type:

STRING

Default:

ALL

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      verify:
        server:
          properties: ALL

Configures Traffic Server for what the default verify callback should check during origin server verification.

You can override this global setting on a per domain basis in the sni.yaml file using the verify_server_properties attribute.

You can also override via the conf_remap plugin. Those changes will take precedence over the changes in sni.yaml.

NONE

Check nothing in the standard callback. Rely entirely on plugins to check the certificate.

SIGNATURE

Check only for a valid signature.

NAME

Check only that the SNI name is in the certificate.

ALL

Check both the signature and the name.

proxy.config.ssl.client.cert.filename

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      cert:
        filename: null

The filename of SSL client certificate installed on Traffic Server.

proxy.config.ssl.client.cert.exit_on_load_fail

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    client:
      cert:
        exit_on_load_fail: 0

By default (0), Traffic Server will start even if problems occur when loading the SSL client certificates. If true (1), SSL client certificate load failures will prevent Traffic Server from starting.

proxy.config.ssl.client.cert.path

Scope:

CONFIG

Type:

STRING

Default:

/config

Reloadable:

Yes

yaml:

records:
  ssl:
    client:
      cert:
        path: /config

The location of the SSL client certificate installed on Traffic Server.

proxy.config.ssl.client.private_key.filename

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      private_key:
        filename: null

The filename of the Traffic Server private key. Change this variable only if the private key is not located in the Traffic Server SSL client certificate file.

proxy.config.ssl.client.private_key.path

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  ssl:
    client:
      private_key:
        path: null

The location of the Traffic Server private key. Change this variable only if the private key is not located in the SSL client certificate file.

proxy.config.ssl.client.CA.cert.filename

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      CA:
        cert:
          filename: null

The filename of the certificate authority against which the origin server will be verified.

proxy.config.ssl.client.CA.cert.path

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  ssl:
    client:
      CA:
        cert:
          path: null

Specifies the location of the certificate authority file against which the origin server will be verified.

proxy.config.ssl.client.sni_policy

Scope:

CONFIG

Type:

STRING

Default:

NULL

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      sni_policy: null

Indicate how the SNI value for the TLS connection to the origin is selected.

host

This is the default. The value of the Host field in the proxy request is used.

server_name

The SNI value of the inbound TLS connection is used.

remap

The remapped upstream name is used.

verify_with_name_source

The value of the Host field in the proxy request is used. In addition, if the names in the server certificate of the upstream are checked, they are checked against the remapped upstream name, not the SNI.

@...

If the policy starts with the @ character, it is treated as a literal, less the leading @. E.g. if the policy is “@apache.org” the SNI will be “apache.org”.

We have two names that could be used in the transaction host header and the SNI value to the origin. These could be the host header from the client or the remap host name. Unless you have pristine host header enabled, these are likely the same values. If sni_policy = host, both the sni and the value of the Host field to origin will be the same. If sni_policy = remap, the sni value will be the remap host name and the host header will be the host header from the client.

In addition, We may want to set the SNI and host headers the same (makes some common web servers happy), but the server certificate for the upstream may have a name that corresponds to the remap name. So instead of using the SNI name for the name check, we may want to use the remap name. So if sni_policy = verify_with_name_source, the sni will be the host header value and the name to check in the server certificate will be the remap header value.

proxy.config.ssl.client.scheme_proto_mismatch_policy

Scope:

CONFIG

Type:

INT

Default:

2

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      scheme_proto_mismatch_policy: 2

This option controls how Traffic Server behaves when the client side connection protocol and the client request’s scheme do not match. For example, if enforcement is enabled by setting this value to 2 and the client connection is a cleartext HTTP connection but the scheme of the URL is https://, then Traffic Server will emit a warning and return an immediate 400 HTTP response without proxying the request to the origin.

The default value is 2, meaning that Traffic Server will enforce that the protocol matches the scheme.

Value	Description
0	Disable verification that the protocol and scheme match.
1	Check that the protocol and scheme match, but only emit a warning if they do not.
2	Check that the protocol and scheme match and, if they do not, emit a warning and return an immediate HTTP 400 response.

proxy.config.ssl.client.version.min

Scope:

CONFIG

Type:

INT

Default:

-1

yaml:

records:
  ssl:
    client:
      version:
        min: -1

Specifies the minimum TLS version that will be offered to servers during the TLS negotiation.

Value	Version
0	TLS 1.0
1	TLS 1.1
2	TLS 1.2
3	TLS 1.3

proxy.config.ssl.client.version.max

Scope:

CONFIG

Type:

INT

Default:

-1

yaml:

records:
  ssl:
    client:
      version:
        max: -1

Specifies the maximum TLS version that will be offered to servers during the TLS negotiation.

Value	Version
0	TLS 1.0
1	TLS 1.1
2	TLS 1.2
3	TLS 1.3

proxy.config.ssl.client.TLSv1

Scope:

CONFIG

Type:

INT

Default:

0

Deprecated:

Yes

yaml:

records:
  ssl:
    client:
      TLSv1: 0

This setting is deprecated in favor of proxy.config.ssl.client.version.min and proxy.config.ssl.client.version.max, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLSv1.0 in the ATS client context. If not specified, disabled by default

proxy.config.ssl.client.TLSv1_1

Scope:

CONFIG

Type:

INT

Default:

0

Deprecated:

Yes

yaml:

records:
  ssl:
    client:
      TLSv1_1: 0

This setting is deprecated in favor of proxy.config.ssl.client.version.min and proxy.config.ssl.client.version.max, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLSv1_1 in the ATS client context. If not specified, disabled by default

proxy.config.ssl.client.TLSv1_2

Scope:

CONFIG

Type:

INT

Default:

1

Deprecated:

Yes

yaml:

records:
  ssl:
    client:
      TLSv1_2: 1

This setting is deprecated in favor of proxy.config.ssl.client.version.min and proxy.config.ssl.client.version.max, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLSv1_2 in the ATS client context. If not specified, enabled by default

proxy.config.ssl.client.TLSv1_3.enabled

Scope:

CONFIG

Type:

INT

Default:

1

Deprecated:

Yes

yaml:

records:
  ssl:
    client:
      TLSv1_3:
        enabled: 1

This setting is deprecated in favor of proxy.config.ssl.client.version.min and proxy.config.ssl.client.version.max, and will be ignored if those new settings are used.

Enables (1) or disables (0) TLSv1_3 in the ATS client context. If not specified, enabled by default

proxy.config.ssl.client.alpn_protocols

Scope:

CONFIG

Type:

STRING

Default:

“”

Overridable:

Yes

yaml:

records:
  ssl:
    client:
      alpn_protocols: '""'

Sets the ALPN string that Traffic Server will send to the origin in the ClientHello of TLS handshakes. Configuring this to an empty string (the default configuration) means that the ALPN extension will not be sent as a part of the TLS ClientHello, resulting in HTTP/1.x being negotiated for all origin-side connections.

Configuring the ALPN string provides a mechanism to control origin-side HTTP protocol negotiation. Including h2 in the ALPN list is required for negotiatnge origin-side HTTP/2 connections.

Configuring this requires an understanding of the ALPN TLS protocol extension. See RFC 7301 for details about the ALPN protocol. See the official IANA ALPN protocol registration for the official list of ALPN protocol names. As a summary, the ALPN string is a comma-separated (no spaces) list of protocol names that the TLS client (Traffic Server in this case) supports. On the TLS server side (origin side in this case), the names are compared in order to the list of protocols supported by the origin. The first match is used, thus the ALPN list should be listed in decreasing order of preference. If no match is found, the TLS server is expected (per the RFC) to fail the TLS handshake with a fatal “no_application_protocol” alert.

Currently, Traffic Server supports the following ALPN protocol names:

• http/1.0

• http/1.1

• h2

Here are some example configurations and the consequences of each:

Value	Description
""	No ALPN extension is sent by Traffic Server in origin-side TLS handshakes. Traffic Server will assume an HTTP/1.1 connection in this case.
"http/1.1"	Only HTTP/1.1 is advertised by Traffic Server. Thus, the origin will either negotiate HTTP/1.1, or it will fail the handshake if that is not supported by the origin.
"http/1.1,http/1.0"	Both HTTP/1.1 and HTTP/1.0 are supported by Traffic Server, but HTTP/1.1 is preferred.
"h2,http/1.1,http/1.0"	HTTP/2 is preferred by Traffic Server over HTTP/1.1 and HTTP/1.0. Thus, if the origin supports HTTP/2, it will be used for the connection. If not, it will fall back to HTTP/1.1 or, if that is not supported, HTTP/1.0. (HTTP/2 to origin is currently not supported by Traffic Server.)
"h2"	Traffic Server only advertises HTTP/2 support. Thus, the origin will either negotiate HTTP/2 or fail the handshake. (HTTP/2 to origin is currently not supported by Traffic Server.)

Note that this is an overridable configuration, so the ALPN can be configured on a per-origin basis via the Configuration Remap Plugin plugin.

proxy.config.ssl.async.handshake.enabled

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    async:
      handshake:
        enabled: 0

Enables the use of OpenSSL async job during the TLS handshake. Traffic Server must be build against OpenSSL 1.1 or greater or this to take affect. Can be useful if using a crypto engine that communicates off chip. The thread will be rescheduled for other work until the crypto engine operation completes. A test crypto engine that inserts a 5 second delay on private key operations can be found at contrib/openssl/async_engine.cc.

proxy.config.ssl.engine.conf_file

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    engine:
      conf_file: null

Specify the location of the OpenSSL config file used to load dynamic crypto engines. This setting assumes an absolute path. An example config file is at contrib/openssl/load_engine.cnf.

TLS v1.3 0-RTT Configuration

Note

TLS v1.3 must be enabled in order to utilize 0-RTT early data.

proxy.config.ssl.server.max_early_data

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    server:
      max_early_data: 0

Specifies the maximum amount of early data in bytes that is permitted to be sent on a single connection.

The minimum value that enables early data, and the suggested value for this option are both 16384 (16KB).

Setting to 0 effectively disables 0-RTT.

If you use BoringSSL, setting a value grater than 0 enables early data but the value won’t be used to limit the maximum amount of early data.

proxy.config.ssl.server.allow_early_data_params

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    server:
      allow_early_data_params: 0

Set to 1 to allow HTTP parameters on early data requests.

SNI Routing

proxy.config.tunnel.activity_check_period

Scope:

CONFIG

Type:

INT

Default:

0

Units:

seconds

yaml:

records:
  tunnel:
    activity_check_period: 0

Frequency of checking the activity of SNI Routing Tunnel. Set to 0 to disable monitoring of the activity of the SNI tunnels. The feature is disabled by default.

proxy.config.tunnel.prewarm.enabled

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  tunnel:
    prewarm:
      enabled: 0

Enable Pre-warming TLS Tunnel. The feature is disabled by default.

proxy.config.tunnel.prewarm.algorithm

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  tunnel:
    prewarm:
      algorithm: 2

Version of pre-warming algorithm.

Value	Description
1	Periodical pre-warming only
2	Event based pre-warming + Periodical pre-warming

proxy.config.tunnel.prewarm.event_period

Scope:

CONFIG

Type:

INT

Default:

1000

Units:

milliseconds

yaml:

records:
  tunnel:
    prewarm:
      event_period: 1000

Frequency of periodical pre-warming in milli-seconds.

OCSP Stapling Configuration

proxy.config.ssl.ocsp.enabled

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    ocsp:
      enabled: 0

Enable OCSP stapling.

Value	Description
0	Disables OCSP Stapling.
1	Allows Traffic Server to request SSL certificate revocation status from an OCSP responder.

proxy.config.ssl.ocsp.cache_timeout

Scope:

CONFIG

Type:

INT

Default:

3600

yaml:

records:
  ssl:
    ocsp:
      cache_timeout: 3600

Number of seconds before an OCSP response expires in the stapling cache.

proxy.config.ssl.ocsp.request_mode

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  ssl:
    ocsp:
      request_mode: 0

Set the request method to prefer when querying OCSP responders. The default is zero, or POST, and a value of 1 will cause ATS to attempt a GET request. Because the length of the encoded request must be less than 255 bytes per RFC 6960, Appendix A, ATS will fall back to the POST request method when the encoded size exceeds this limit.

proxy.config.ssl.ocsp.request_timeout

Scope:

CONFIG

Type:

INT

Default:

10

Units:

seconds

yaml:

records:
  ssl:
    ocsp:
      request_timeout: 10

Timeout (in seconds) for queries to OCSP responders.

proxy.config.ssl.ocsp.update_period

Scope:

CONFIG

Type:

INT

Default:

60

Units:

seconds

yaml:

records:
  ssl:
    ocsp:
      update_period: 60

Update period (in seconds) for stapling caches.

proxy.config.ssl.ocsp.response.path

Scope:

CONFIG

Type:

STRING

Default:

NULL

yaml:

records:
  ssl:
    ocsp:
      response:
        path: null

The directory path of the prefetched OCSP stapling responses. Change this variable only if you intend to use and administratively maintain prefetched OCSP stapling responses. All stapling responses listed in ssl_multicert.config will be loaded relative to this path.

HTTP/2 Configuration

proxy.config.http2.max_concurrent_streams_in

Scope:

CONFIG

Type:

INT

Default:

100

Reloadable:

Yes

yaml:

records:
  http2:
    max_concurrent_streams_in: 100

The maximum number of concurrent streams per inbound connection.

Note

Reloading this value affects only new HTTP/2 connections, not the ones already established.

proxy.config.http2.max_concurrent_streams_out

Scope:

CONFIG

Type:

INT

Default:

100

Reloadable:

Yes

yaml:

records:
  http2:
    max_concurrent_streams_out: 100

The maximum number of concurrent streams per outbound connection.

Note

Reloading this value affects only new HTTP/2 connections, not the ones already established.

proxy.config.http2.min_concurrent_streams_in

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

yaml:

records:
  http2:
    min_concurrent_streams_in: 10

The minimum number of concurrent streams per inbound connection. This is used when proxy.config.http2.max_active_streams_in is set larger than 0.

proxy.config.http2.min_concurrent_streams_out

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

yaml:

records:
  http2:
    min_concurrent_streams_out: 10

The minimum number of concurrent streams per outbound connection. This is used when proxy.config.http2.max_active_streams_out is set larger than 0.

proxy.config.http2.max_active_streams_in

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http2:
    max_active_streams_in: 0

Limits the maximum number of connection wide active streams. When connection wide active streams are larger than this value, SETTINGS_MAX_CONCURRENT_STREAMS will be reduced to proxy.config.http2.min_concurrent_streams_in. To disable, set to zero (0).

proxy.config.http2.max_active_streams_out

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http2:
    max_active_streams_out: 0

Limits the maximum number of connection wide active streams. When connection wide active streams are larger than this value, SETTINGS_MAX_CONCURRENT_STREAMS will be reduced to proxy.config.http2.min_concurrent_streams_out. To disable, set to zero (0).

proxy.config.http2.initial_window_size_in

Scope:

CONFIG

Type:

INT

Default:

65535

Units:

bytes

Reloadable:

Yes

yaml:

records:
  http2:
    initial_window_size_in: 65535

The initial HTTP/2 stream window size for inbound connections that Traffic Server as a receiver advertises to the peer. See IETF RFC 9113 section 5.2 for details concerning HTTP/2 flow control. See proxy.config.http2.flow_control.policy_in for how HTTP/2 stream and session windows are maintained over the lifetime of HTTP/2 sessions.

proxy.config.http2.initial_window_size_out

Scope:

CONFIG

Type:

INT

Default:

65535

Units:

bytes

Reloadable:

Yes

yaml:

records:
  http2:
    initial_window_size_out: 65535

The initial HTTP/2 stream window size for outbound connections that Traffic Server as a client advertises to the peer. See IETF RFC 9113 section 5.2 for details concerning HTTP/2 flow control. See proxy.config.http2.flow_control.policy_out for how HTTP/2 stream and session windows are maintained over the lifetime of HTTP/2 sessions.

proxy.config.http2.flow_control.policy_in

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http2:
    flow_control:
      policy_in: 0

Specifies the mechanism Traffic Server uses to maintian flow control via the HTTP/2 stream and session windows for inbound connections. See IETF RFC 9113 section 5.2 for details concerning HTTP/2 flow control.

Value	Description
0	Session and stream receive windows are initialized and maintained at the value as specified in proxy.config.http2.initial_window_size_in over the lifetime of HTTP/2 sessions.
1	Session receive windows are initialized to the value of the product of proxy.config.http2.initial_window_size_in and proxy.config.http2.max_concurrent_streams_in and are maintained as such over the lifetime of HTTP/2 sessions. Stream windows are initialized to the value of proxy.config.http2.initial_window_size_in and are maintained as such over the lifetime of each HTTP/2 stream.
2	Session receive windows are initialized to the value of the product of proxy.config.http2.initial_window_size_in and proxy.config.http2.max_concurrent_streams_in and are maintained as such over the lifetime of HTTP/2 sessions. Stream windows are initialized to the value of proxy.config.http2.initial_window_size_in but are dynamically adjusted to the session window size divided by the number of concurrent streams over the lifetime of HTTP/2 sessions. That is, stream window sizes dynamically adjust to fill the session window in a way that shares the window equally among all concurrent streams.

proxy.config.http2.flow_control.policy_out

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http2:
    flow_control:
      policy_out: 0

Specifies the mechanism Traffic Server uses to maintian flow control via the HTTP/2 stream and session windows for outbound connections. See the corresponding proxy.config.http2.flow_control.policy_in configuration for details concerning how this configuration variable is used.

proxy.config.http2.max_frame_size

Scope:

CONFIG

Type:

INT

Default:

16384

Units:

bytes

Reloadable:

Yes

yaml:

records:
  http2:
    max_frame_size: 16384

Indicates the size of the largest frame payload that the sender is willing to receive.

proxy.config.http2.header_table_size

Scope:

CONFIG

Type:

INT

Default:

4096

Reloadable:

Yes

yaml:

records:
  http2:
    header_table_size: 4096

The maximum size of the header compression table used to decode header blocks. This value will be advertised as SETTINGS_HEADER_TABLE_SIZE.

proxy.config.http2.header_table_size_limit

Scope:

CONFIG

Type:

INT

Default:

65536

Reloadable:

Yes

yaml:

records:
  http2:
    header_table_size_limit: 65536

The maximum size of the header compression table ATS actually use when ATS encodes headers. Setting 0 means ATS doesn’t insert headers into HPACK Dynamic Table, however, headers still can be encoded as indexable representations. The upper limit is 65536.

proxy.config.http2.max_header_list_size

Scope:

CONFIG

Type:

INT

Default:

32768

Reloadable:

Yes

yaml:

records:
  http2:
    max_header_list_size: 32768

This advisory setting informs a peer of the maximum size of header list that the sender is prepared to accept.

proxy.config.http2.stream_priority_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http2:
    stream_priority_enabled: 0

Enable the experimental HTTP/2 Stream Priority feature.

proxy.config.http2.active_timeout_in

Scope:

CONFIG

Type:

INT

Default:

0

Units:

seconds

Reloadable:

Yes

yaml:

records:
  http2:
    active_timeout_in: 0

This is the active timeout of the http2 connection. It is set when the connection is opened and keeps ticking regardless of activity level.

The value of 0 specifies that there is no timeout.

proxy.config.http2.accept_no_activity_timeout

Scope:

CONFIG

Type:

INT

Default:

120

Units:

seconds

Reloadable:

Yes

yaml:

records:
  http2:
    accept_no_activity_timeout: 120

Specifies how long Traffic Server keeps connections to clients open if no activity is received on the connection. Lowering this timeout can ease pressure on the proxy if misconfigured or misbehaving clients are opening a large number of connections without submitting requests.

proxy.config.http2.no_activity_timeout_in

Scope:

CONFIG

Type:

INT

Default:

120

Units:

seconds

Reloadable:

Yes

yaml:

records:
  http2:
    no_activity_timeout_in: 120

Specifies how long Traffic Server keeps connections to clients open if a transaction stalls. Lowering this timeout can ease pressure on the proxy if misconfigured or misbehaving clients are opening a large number of connections without submitting requests.

proxy.config.http2.no_activity_timeout_out

Scope:

CONFIG

Type:

INT

Default:

120

Units:

seconds

Reloadable:

Yes

yaml:

records:
  http2:
    no_activity_timeout_out: 120

Specifies how long Traffic Server keeps connections to origins open if a transaction stalls.

proxy.config.http2.incomplete_header_timeout_in

Scope:

CONFIG

Type:

INT

Default:

10

Units:

seconds

Reloadable:

Yes

yaml:

records:
  http2:
    incomplete_header_timeout_in: 10

Specifies how long Traffic Server keeps streams to clients open after they start sending HTTP headers. If a client doesn’t send all headers within this time, the stream and connection will be closed.

proxy.config.http2.zombie_debug_timeout_in

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http2:
    zombie_debug_timeout_in: 0

This timeout enables the zombie debugging feature. If it is non-zero, it sets a zombie event to go off that many seconds in the future when the HTTP2 session reaches one but not both of the terminating events, i.e received a close event (via client goaway or timeout) and the number of active streams has gone to zero. If the event is executed, the Traffic Server process will assert. This mechanism is useful to debug potential leaks in the HTTP2 Stream and Session processing.

proxy.config.http2.push_diary_size

Scope:

CONFIG

Type:

INT

Default:

256

Reloadable:

Yes

yaml:

records:
  http2:
    push_diary_size: 256

Indicates the maximum number of HTTP/2 server pushes that are remembered per HTTP/2 connection to avoid duplicate pushes on the same connection. If the maximum number is reached, new entries are not remembered.

proxy.config.http2.stream_error_rate_threshold

Scope:

CONFIG

Type:

FLOAT

Default:

0.1

Reloadable:

Yes

yaml:

records:
  http2:
    stream_error_rate_threshold: 0.1

This is the maximum stream error rate Traffic Server allows on an HTTP/2 connection. Traffic Server gracefully closes connections that have stream error rates above this setting by sending GOAWAY frames.

proxy.config.http2.stream_error_sampling_threshold

Scope:

CONFIG

Type:

INT

Default:

10

Reloadable:

Yes

yaml:

records:
  http2:
    stream_error_sampling_threshold: 10

This is the threshold of sampling stream number to start checking the stream error rate.

proxy.config.http2.max_settings_per_frame

Scope:

CONFIG

Type:

INT

Default:

7

Reloadable:

Yes

yaml:

records:
  http2:
    max_settings_per_frame: 7

Specifies how many settings in an HTTP/2 SETTINGS frame Traffic Server accepts. Clients exceeded this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. Any negative value configures no limit to the number of settings received.

proxy.config.http2.max_settings_per_minute

Scope:

CONFIG

Type:

INT

Default:

14

Reloadable:

Yes

yaml:

records:
  http2:
    max_settings_per_minute: 14

Specifies how many settings in HTTP/2 SETTINGS frames Traffic Server accept for a minute. Clients exceeded this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. Any negative value configures no limit to the number of settings received.

proxy.config.http2.max_settings_frames_per_minute

Scope:

CONFIG

Type:

INT

Default:

14

Reloadable:

Yes

yaml:

records:
  http2:
    max_settings_frames_per_minute: 14

Specifies how many SETTINGS frames Traffic Server receives for a minute at maximum. Clients exceeded this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. Any negative value configures no limit to the number of SETTINGS frames received.

proxy.config.http2.max_ping_frames_per_minute

Scope:

CONFIG

Type:

INT

Default:

60

Reloadable:

Yes

yaml:

records:
  http2:
    max_ping_frames_per_minute: 60

Specifies how many number of PING frames Traffic Server receives for a minute at maximum. Clients exceeded this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. Any negative value configures no limit to the number of PING frames received.

proxy.config.http2.max_priority_frames_per_minute

Scope:

CONFIG

Type:

INT

Default:

120

Reloadable:

Yes

yaml:

records:
  http2:
    max_priority_frames_per_minute: 120

Specifies how many number of PRIORITY frames Traffic Server receives for a minute at maximum. Clients exceeded this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. If this is set to 0, the limit logic is disabled. This limit only will be enforced if proxy.config.http2.stream_priority_enabled is set to 1. Any negative value configures no limit to the number of PRIORITY frames received.

proxy.config.http2.max_rst_stream_frames_per_minute

Scope:

CONFIG

Type:

INT

Default:

200

Reloadable:

Yes

yaml:

records:
  http2:
    max_rst_stream_frames_per_minute: 200

Specifies how many RST_STREAM frames Traffic Server receives per minute at maximum. Clients exceeding this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. Any negative value configures no limit to the number of RST_STREAM frames received.

proxy.config.http2.max_continuation_frames_per_minute

Scope:

CONFIG

Type:

INT

Default:

120

Reloadable:

Yes

yaml:

records:
  http2:
    max_continuation_frames_per_minute: 120

Specifies how many CONTINUATION frames Traffic Server receives per minute at maximum. Clients exceeding this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. Any negative value configures no limit to the number of CONTINUATION frames received.

proxy.config.http2.max_empty_frames_per_minute

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  http2:
    max_empty_frames_per_minute: 0

Specifies the maximum number of empty frames Traffic Server will receive per minute before it will start closing connections. In this context, an “empty frame” means either a DATA frame that does not carry a payload nor an END_STREAM flag, or a CONTINUATION frame that does not carry payload nor an END_HEADERS flag. Clients exceeding this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM. Any negative value configures no limit to the number of empty frames received. 0 is the default configuration, meaning that no empty frames are allowed.

proxy.config.http2.min_avg_window_update

Scope:

CONFIG

Type:

FLOAT

Default:

2560.0

Reloadable:

Yes

yaml:

records:
  http2:
    min_avg_window_update: 2560.0

Specifies the minimum average window increment Traffic Server allows. The average will be calculated based on the last 5 WINDOW_UPDATE frames. Clients that send smaller window increments lower than this limit will be immediately disconnected with an error code of ENHANCE_YOUR_CALM.

proxy.config.http2.write_buffer_block_size

Scope:

CONFIG

Type:

INT

Default:

262144

Units:

bytes

Reloadable:

Yes

yaml:

records:
  http2:
    write_buffer_block_size: 262144

Specifies the size of a buffer block that is used for buffering outgoing HTTP/2 frames. The size will be rounded up based on power of 2.

proxy.config.http2.write_size_threshold

Scope:

CONFIG

Type:

FLOAT

Default:

0.5

Reloadable:

Yes

yaml:

records:
  http2:
    write_size_threshold: 0.5

Specifies the size threshold for triggering write operation for sending HTTP/2 frames. The default value is 0.5 and it measn write operation is going to be triggered when half or more of the buffer is occupied.

proxy.config.http2.write_time_threshold

Scope:

CONFIG

Type:

INT

Default:

100

Units:

milliseconds

Reloadable:

Yes

yaml:

records:
  http2:
    write_time_threshold: 100

Specifies the time threshold for triggering write operation for sending HTTP/2 frames. Write operation will be triggered at least once every this configured number of millisecond regardless of pending data size.

proxy.config.http2.default_buffer_water_mark

Scope:

CONFIG

Type:

INT

Default:

-1

Units:

bytes

Reloadable:

Yes

yaml:

records:
  http2:
    default_buffer_water_mark: -1

Specifies the high water mark for all HTTP/2 frames on an outgoing connection. Default is -1 to preserve existing water marking behavior.

You can override this global setting on a per domain basis in the sni.yaml file using the http2_buffer_water_mark attribute.

HTTP/3 Configuration

There is no configuration available yet on this release.

QUIC Configuration

All configurations for QUIC are still experimental and may be changed or removed in the future without prior notice.

proxy.config.quic.qlog.file_base

Scope:

CONFIG

Type:

STRING

Default:

NULL

Reloadable:

Yes

yaml:

records:
  quic:
    qlog:
      file_base: null

Sets qlog output to a specific base file name. For a given file base name the file becomes {file_base}-{trace id}.sqlog. Absolute path or relative to the configured prefix can be used. Qlog is emitted for each connection.

records:
  quic:
    qlog:
      file_base: /my/logs/test1

The above configuration will make Traffic Server to generate Qlogs with the following format:

/my/logs/test1-e9158839e54cb8f34ab00d34236ad5e19a49.sqlog

proxy.config.quic.instance_id

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  quic:
    instance_id: 0

A static key used for calculating Stateless Reset Token. All instances in a cluster need to share the same value.

proxy.config.quic.connection_table.size

Scope:

CONFIG

Type:

INT

Default:

65521

yaml:

records:
  quic:
    connection_table:
      size: 65521

A size of hash table that stores connection information.

proxy.config.quic.proxy.config.quic.num_alt_connection_ids

Scope:

CONFIG

Type:

INT

Default:

65521

Reloadable:

Yes

yaml:

records:
  quic:
    proxy:
      config:
        quic:
          num_alt_connection_ids: 65521

A number of alternate Connection IDs that Traffic Server provides to a peer. It has to be at least 8.

proxy.config.quic.stateless_retry_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  quic:
    stateless_retry_enabled: 0

Enables Stateless Retry.

proxy.config.quic.client.vn_exercise_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  quic:
    client:
      vn_exercise_enabled: 0

Enables version negotiation exercise on origin server connections.

proxy.config.quic.client.cm_exercise_enabled

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  quic:
    client:
      cm_exercise_enabled: 0

Enables connection migration exercise on origin server connections.

proxy.config.quic.server.supported_groups

Scope:

CONFIG

Type:

STRING

Default:

“P-256:X25519:P-384:P-521”

Reloadable:

Yes

yaml:

records:
  quic:
    server:
      supported_groups: '"P-256:X25519:P-384:P-521"'

Configures the list of supported groups provided by OpenSSL which will be used to determine the set of shared groups on QUIC origin server connections.

proxy.config.quic.client.supported_groups

Scope:

CONFIG

Type:

STRING

Default:

“P-256:X25519:P-384:P-521”

Reloadable:

Yes

yaml:

records:
  quic:
    client:
      supported_groups: '"P-256:X25519:P-384:P-521"'

Configures the list of supported groups provided by OpenSSL which will be used to determine the set of shared groups on QUIC client connections.

proxy.config.quic.client.session_file

Scope:

CONFIG

Type:

STRING

Default:

“”

Reloadable:

Yes

yaml:

records:
  quic:
    client:
      session_file: '""'

Only available for traffic_quic. If specified, TLS session data will be stored to the file, and will be used for resuming a session.

proxy.config.quic.no_activity_timeout_in

Scope:

CONFIG

Type:

INT

Default:

30000

Reloadable:

Yes

yaml:

records:
  quic:
    no_activity_timeout_in: 30000

The max idle timeout in milliseconds. This value will be advertised as idle_timeout Transport Parameter. Transport Parameter.

Important

QUIC ignores any settings for proxy.config.net.default_inactivity_timeout and only honors the proxy.config.quic.no_activity_timeout_in.

proxy.config.quic.no_activity_timeout_out

Scope:

CONFIG

Type:

INT

Default:

30000

Reloadable:

Yes

yaml:

records:
  quic:
    no_activity_timeout_out: 30000

The max idle timeout in milliseconds. This value will be advertised as idle_timeout Transport Parameter. Transport Parameter.

proxy.config.quic.preferred_address_ipv4

Scope:

CONFIG

Type:

STRING

Default:

“”

Reloadable:

Yes

yaml:

records:
  quic:
    preferred_address_ipv4: '""'

This value will be advertised as a part of preferred_address Transport Parameter.

proxy.config.quic.preferred_address_ipv6

Scope:

CONFIG

Type:

STRING

Default:

“”

Reloadable:

Yes

yaml:

records:
  quic:
    preferred_address_ipv6: '""'

This value will be advertised as a part of preferred_address Transport Parameter.

proxy.config.quic.initial_max_data_in

Scope:

CONFIG

Type:

INT

Default:

65536

Reloadable:

Yes

yaml:

records:
  quic:
    initial_max_data_in: 65536

Integer value that contains the initial value for the maximum amount of data that can be sent on the connection.

This value will be advertised as initial_max_data Transport Parameter.

proxy.config.quic.initial_max_data_out

Scope:

CONFIG

Type:

INT

Default:

65536

Reloadable:

Yes

yaml:

records:
  quic:
    initial_max_data_out: 65536

This value will be advertised as initial_max_data Transport Parameter.

proxy.config.quic.max_stream_data_bidi_local_in

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  quic:
    max_stream_data_bidi_local_in: 0

This value will be advertised as initial_max_stream_data_bidi_local Transport Parameter.

proxy.config.quic.max_stream_data_bidi_local_out

Scope:

CONFIG

Type:

INT

Default:

4096

Reloadable:

Yes

yaml:

records:
  quic:
    max_stream_data_bidi_local_out: 4096

This value will be advertised as initial_max_stream_data_bidi_local Transport Parameter.

proxy.config.quic.max_stream_data_bidi_remote_in

Scope:

CONFIG

Type:

INT

Default:

4096

Reloadable:

Yes

yaml:

records:
  quic:
    max_stream_data_bidi_remote_in: 4096

This value will be advertised as initial_max_stream_data_bidi_remote Transport Parameter.

proxy.config.quic.max_stream_data_bidi_remote_out

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  quic:
    max_stream_data_bidi_remote_out: 0

This value will be advertised as initial_max_stream_data_bidi_remote Transport Parameter.

proxy.config.quic.max_stream_data_uni_in

Scope:

CONFIG

Type:

INT

Default:

4096

Reloadable:

Yes

yaml:

records:
  quic:
    max_stream_data_uni_in: 4096

This value will be advertised as initial_max_stream_data_uni Transport Parameter.

proxy.config.quic.max_stream_data_uni_out

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  quic:
    max_stream_data_uni_out: 0

This value will be advertised as initial_max_stream_data_uni Transport Parameter.

proxy.config.quic.max_streams_bidi_in

Scope:

CONFIG

Type:

INT

Default:

100

Reloadable:

Yes

yaml:

records:
  quic:
    max_streams_bidi_in: 100

This value will be advertised as initial_max_streams_bidi Transport Parameter.

proxy.config.quic.max_streams_bidi_out

Scope:

CONFIG

Type:

INT

Default:

100

Reloadable:

Yes

yaml:

records:
  quic:
    max_streams_bidi_out: 100

This value will be advertised as initial_max_streams_bidi Transport Parameter.

proxy.config.quic.max_streams_uni_in

Scope:

CONFIG

Type:

INT

Default:

100

Reloadable:

Yes

yaml:

records:
  quic:
    max_streams_uni_in: 100

This value will be advertised as initial_max_streams_uni Transport Parameter.

proxy.config.quic.max_streams_uni_out

Scope:

CONFIG

Type:

INT

Default:

100

Reloadable:

Yes

yaml:

records:
  quic:
    max_streams_uni_out: 100

This value will be advertised as initial_max_streams_uni Transport Parameter.

proxy.config.quic.ack_delay_exponent_in

Scope:

CONFIG

Type:

INT

Default:

3

Reloadable:

Yes

yaml:

records:
  quic:
    ack_delay_exponent_in: 3

This value will be advertised as ack_delay_exponent Transport Parameter.

proxy.config.quic.ack_delay_exponent_out

Scope:

CONFIG

Type:

INT

Default:

3

Reloadable:

Yes

yaml:

records:
  quic:
    ack_delay_exponent_out: 3

This value will be advertised as ack_delay_exponent Transport Parameter.

proxy.config.quic.max_ack_delay_in

Scope:

CONFIG

Type:

INT

Default:

25

Reloadable:

Yes

yaml:

records:
  quic:
    max_ack_delay_in: 25

This value will be advertised as max_ack_delay Transport Parameter.

proxy.config.quic.max_ack_delay_out

Scope:

CONFIG

Type:

INT

Default:

25

Reloadable:

Yes

yaml:

records:
  quic:
    max_ack_delay_out: 25

This value will be advertised as max_ack_delay Transport Parameter.

proxy.config.quic.active_cid_limit_in

Scope:

CONFIG

Type:

INT

Default:

2

Reloadable:

Yes

yaml:

records:
  quic:
    active_cid_limit_in: 2

Integer value specifying the maximum number of connection IDs from the peer that Traffic Server is willing to store. The value MUST be at least 2. Transport Parameter.

proxy.config.quic.max_recv_udp_payload_size_in

Scope:

CONFIG

Type:

INT

Default:

65527

yaml:

records:
  quic:
    max_recv_udp_payload_size_in: 65527

This value will be advertised as max_udp_payload_size Transport Parameter.

proxy.config.quic.max_recv_udp_payload_size_out

Scope:

CONFIG

Type:

INT

Default:

65527

yaml:

records:
  quic:
    max_recv_udp_payload_size_out: 65527

This value will be advertised as max_udp_payload_size Transport Parameter.

proxy.config.quic.max_send_udp_payload_size_in

Scope:

CONFIG

Type:

INT

Default:

65527

yaml:

records:
  quic:
    max_send_udp_payload_size_in: 65527

Specified the maximum outgoing UDP payload size.

proxy.config.quic.max_send_udp_payload_size_out

Scope:

CONFIG

Type:

INT

Default:

65527

yaml:

records:
  quic:
    max_send_udp_payload_size_out: 65527

Specified the maximum outgoing UDP payload size.

proxy.config.quic.disable_http_0_9

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  quic:
    disable_http_0_9: 1

Disables HTTP/0.9 over QUIC by default.

proxy.config.quic.cc_algorithm

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  quic:
    cc_algorithm: 0

Specified the congestion control algorithm.

Value	Description
0	RENO (default).
1	CUBIC.

UDP Configuration

proxy.config.udp.poll_timeout

Scope:

CONFIG

Type:

INT

Default:

100

Units:

milliseconds

yaml:

records:
  udp:
    poll_timeout: 100

This is the timeout for listening UDP connections to epoll_wait() on Linux platforms, and to kevent() on BSD type OSs. The default value is 100. See proxy.config.net.poll_timeout for general information on poll_timeout.

proxy.config.udp.threads

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  udp:
    threads: 0

Specifies the number of UDP threads to run. By default 0 threads are dedicated to UDP, which results in effectively disabling UDP support.

proxy.config.udp.enable_gso

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  udp:
    enable_gso: 1

Enables (1) or disables (0) UDP GSO. When enabled, Traffic Server tries to use UDP GSO, and disables it automatically if it causes send errors.

proxy.config.udp.enable_gro

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  udp:
    enable_gro: 1

Enables (1) or disables (0) UDP GRO. When enabled, Traffic Server will try to use it when reading the UDP socket.

Plug-in Configuration

proxy.config.plugin.plugin_dir

Scope:

CONFIG

Type:

STRING

Default:

config/plugins

yaml:

records:
  plugin:
    plugin_dir: config/plugins

Specifies the location of Traffic Server plugins.

proxy.config.plugin.dynamic_reload_mode

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  plugin:
    dynamic_reload_mode: 1

Enables (1) or disables (0) the dynamic reload feature for remap plugins (remap.config). Global plugins (plugin.config) do not have dynamic reload feature yet.

proxy.config.plugin.compiler_path

Scope:

CONFIG

Type:

STRING

Default:

“”

yaml:

records:
  plugin:
    compiler_path: '""'

Specifies an optional compiler tool path for compiling plugins. This tool should be an executable, which takes two arguments:

Arg	Description
1	This is the path to the source file, which should be compiled
2	This is the path to the DSO file, which will be created and loaded

The script should exit with a status code of 0 if the compilation was successful.

proxy.config.plugin.vc.default_buffer_index

Scope:

CONFIG

Type:

INT

Default:

8

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  plugin:
    vc:
      default_buffer_index: 8

Specifies the buffer index and thus size to use when constructing IO buffers within the PluginVC. Tuning this can impact performance of intercept plugins. Default is 8, which aligns with the default value of ts:cv:CONFIG proxy.config.http.default_buffer_size.

proxy.config.plugin.vc.default_buffer_water_mark

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  plugin:
    vc:
      default_buffer_water_mark: 0

Specifies the buffer water mark size in bytes used to control the flow of data through IO buffers within the PluginVC. Default is zero to preserve existing PluginVC water marking behavior.

SOCKS Processor

proxy.config.socks.socks_needed

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  socks:
    socks_needed: 0

Enables (1) or disables (0) the SOCKS processor

proxy.config.socks.socks_version

Scope:

CONFIG

Type:

INT

Default:

4

yaml:

records:
  socks:
    socks_version: 4

Specifies the SOCKS version (4) or (5)

proxy.config.socks.socks_config_file

Scope:

CONFIG

Type:

STRING

Default:

socks.config

Deprecated:

Yes

yaml:

records:
  socks:
    socks_config_file: socks.config

The socks.config file allows you to specify ranges of IP addresses that will not be relayed to the SOCKS server. It can also be used to configure AUTH information for SOCKSv5 servers.

proxy.config.socks.socks_timeout

Scope:

CONFIG

Type:

INT

Default:

100

yaml:

records:
  socks:
    socks_timeout: 100

The activity timeout value (in seconds) for SOCKS server connections.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.socks.server_connect_timeout

Scope:

CONFIG

Type:

INT

Default:

10

yaml:

records:
  socks:
    server_connect_timeout: 10

The timeout value (in seconds) for SOCKS server connection attempts.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.socks.per_server_connection_attempts

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  socks:
    per_server_connection_attempts: 1

The total number of connection attempts allowed per SOCKS server, if multiple servers are used.

proxy.config.socks.connection_attempts

Scope:

CONFIG

Type:

INT

Default:

4

yaml:

records:
  socks:
    connection_attempts: 4

The total number of connection attempts allowed to a SOCKS server Traffic Server bypasses the server or fails the request

proxy.config.socks.server_retry_timeout

Scope:

CONFIG

Type:

INT

Default:

300

yaml:

records:
  socks:
    server_retry_timeout: 300

The timeout value (in seconds) for SOCKS server connection retry attempts.

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.socks.default_servers

Scope:

CONFIG

Type:

STRING

Default:

“”

yaml:

records:
  socks:
    default_servers: '""'

Default list of SOCKS servers and their ports.

proxy.config.socks.server_retry_time

Scope:

CONFIG

Type:

INT

Default:

300

yaml:

records:
  socks:
    server_retry_time: 300

The amount of time allowed between connection retries to a SOCKS server that is unavailable.

proxy.config.socks.server_fail_threshold

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  socks:
    server_fail_threshold: 2

The number of times the connection to the SOCKS server can fail before Traffic Server considers the server unavailable.

proxy.config.socks.accept_enabled

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  socks:
    accept_enabled: 0

Enables (1) or disables (0) the SOCKS proxy option. As a SOCKS proxy, Traffic Server receives SOCKS traffic (usually on port 1) and forwards all requests directly to the SOCKS server.

proxy.config.socks.accept_port

Scope:

CONFIG

Type:

INT

Default:

1080

yaml:

records:
  socks:
    accept_port: 1080

Specifies the port on which Traffic Server accepts SOCKS traffic.

proxy.config.socks.http_port

Scope:

CONFIG

Type:

INT

Default:

80

yaml:

records:
  socks:
    http_port: 80

Specifies the port on which Traffic Server accepts HTTP proxy requests over SOCKS connections..

Sockets

proxy.config.net.defer_accept

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  net:
    defer_accept: 1

default: 1 meaning on all Platforms except Linux: 45 seconds

This directive enables operating system specific optimizations for a listening socket. defer_accept holds a call to accept(2) back until data has arrived. In Linux’ special case this is up to a maximum of 45 seconds. On FreeBSD, accf_data module needs to be loaded. Note: If MPTCP is enabled, TCP_DEFER_ACCEPT is only supported on Linux kernels 5.19+.

proxy.config.net.listen_backlog

Scope:

CONFIG

Type:

INT

Default:

-1

Reloadable:

Yes

yaml:

records:
  net:
    listen_backlog: -1

This directive sets the maximum number of pending connections. If it is set to -1, Traffic Server will automatically set this to a platform-specific maximum.

proxy.config.net.tcp_congestion_control_in

Scope:

CONFIG

Type:

STRING

Default:

“”

yaml:

records:
  net:
    tcp_congestion_control_in: '""'

This directive will override the congestion control algorithm for incoming connections (accept sockets). On Linux, the allowed values are typically specified in a space separated list in /proc/sys/net/ipv4/tcp_allowed_congestion_control

proxy.config.net.tcp_congestion_control_out

Scope:

CONFIG

Type:

STRING

Default:

“”

yaml:

records:
  net:
    tcp_congestion_control_out: '""'

This directive will override the congestion control algorithm for outgoing connections (connect sockets). On Linux, the allowed values are typically specified in a space separated list in /proc/sys/net/ipv4/tcp_allowed_congestion_control

proxy.config.net.sock_send_buffer_size_in

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  net:
    sock_send_buffer_size_in: 0

Sets the send buffer size for connections from the client to Traffic Server.

proxy.config.net.sock_recv_buffer_size_in

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  net:
    sock_recv_buffer_size_in: 0

Sets the receive buffer size for connections from the client to Traffic Server.

proxy.config.net.sock_option_flag_in

Scope:

CONFIG

Type:

INT

Default:

0x1

yaml:

records:
  net:
    sock_option_flag_in: '0x1'

Turns different options “on” for the socket handling client connections::

TCP_NODELAY  (1)
SO_KEEPALIVE (2)
SO_LINGER (4) - with a timeout of 0 seconds
TCP_FASTOPEN (8)
PACKET_MARK (16)
PACKET_TOS (32)
TCP_NOTSENT_LOWAT (64)
INCOMING_CPU (128)

Note: If MPTCP is enabled, TCP_NODELAY is only supported on Linux kernels 5.17+ and TCP_FASTOPEN is only supported on Linux kernels 6.2+. TCP_NOTSENT_LOWAT socket option is currently not supported.

Note

This is a bitmask and you need to decide what bits to set. Therefore, you must set the value to 3 if you want to enable nodelay and keepalive options above.

Note

To allow TCP Fast Open for client sockets on Linux, bit 2 of the net.ipv4.tcp_fastopen sysctl must be set.

Note

As for SO_INCOMING_CPU, using it with SO_REUSEPORT and exec_thread affinity 4 is recommended.

records:
  accept_threads: 0
  exec_thread:
    listen: 1
    affinity: 4
  net:
   sock_option_flag_in: 0x80

proxy.config.net.sock_send_buffer_size_out

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  net:
    sock_send_buffer_size_out: 0

Sets the send buffer size for connections from Traffic Server to the origin server.

proxy.config.net.sock_recv_buffer_size_out

Scope:

CONFIG

Type:

INT

Default:

0

Overridable:

Yes

yaml:

records:
  net:
    sock_recv_buffer_size_out: 0

Sets the receive buffer size for connections from Traffic Server to the origin server.

proxy.config.net.sock_option_flag_out

Scope:

CONFIG

Type:

INT

Default:

0x1

Overridable:

Yes

yaml:

records:
  net:
    sock_option_flag_out: '0x1'

Turns different options “on” for the origin server socket::

TCP_NODELAY  (1)
SO_KEEPALIVE (2)
SO_LINGER (4) - with a timeout of 0 seconds
TCP_FASTOPEN (8)
PACKET_MARK (16)
PACKET_TOS (32)
TCP_NOTSENT_LOWAT (64)

Note

This is a bitmask and you need to decide what bits to set. Therefore, you must set the value to 3 if you want to enable nodelay and keepalive options above.

When SO_LINGER is enabled, the linger timeout time is set to 0. This is useful when Traffic Server and the origin server are co-located and large numbers of sockets are retained in the TIME_WAIT state.

Note

To allow TCP Fast Open for server sockets on Linux, bit 1 of the net.ipv4.tcp_fastopen sysctl must be set.

proxy.config.net.sock_mss_in

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  net:
    sock_mss_in: 0

Same as the command line option --accept_mss that sets the MSS for all incoming requests. Note: If MPTCP is enabled, TCP_MAXSEG socket option is not supported.

proxy.config.net.sock_packet_mark_in

Scope:

CONFIG

Type:

INT

Default:

0x0

yaml:

records:
  net:
    sock_packet_mark_in: '0x0'

Set the packet mark on traffic destined for the client (the packets that make up a client response).

See also

Traffic Shaping

proxy.config.net.sock_packet_mark_out

Scope:

CONFIG

Type:

INT

Default:

0x0

Overridable:

Yes

yaml:

records:
  net:
    sock_packet_mark_out: '0x0'

Set the packet mark on traffic destined for the origin (the packets that make up an origin request).

See also

Traffic Shaping

proxy.config.net.sock_packet_tos_in

Scope:

CONFIG

Type:

INT

Default:

0x0

yaml:

records:
  net:
    sock_packet_tos_in: '0x0'

Set the ToS/DiffServ Field on packets sent to the client (the packets that make up a client response).

See also

Traffic Shaping

proxy.config.net.sock_packet_tos_out

Scope:

CONFIG

Type:

INT

Default:

0x0

Overridable:

Yes

yaml:

records:
  net:
    sock_packet_tos_out: '0x0'

Set the ToS/DiffServ Field on packets sent to the origin (the packets that make up an origin request).

See also

Traffic Shaping

proxy.config.net.sock_notsent_lowat

Scope:

CONFIG

Type:

INT

Default:

16384

Overridable:

Yes

yaml:

records:
  net:
    sock_notsent_lowat: 16384

Set socket option TCP_NOTSENT_LOWAT to specified value for a connection

proxy.config.net.poll_timeout

Scope:

CONFIG

Type:

INT

Default:

10

yaml:

records:
  net:
    poll_timeout: 10

Same as the command line option --poll_timeout, or -t, which specifies the timeout used for the polling mechanism used. This timeout is always in milliseconds (ms). This is the timeout to epoll_wait() on Linux platforms, and to kevent() on BSD type OSs. The default value is 10 on all platforms.

Changing this configuration can reduce CPU usage on an idle system, since periodic tasks gets processed at these intervals. On busy servers, this overhead is diminished, since polled events triggers more frequently. However, increasing the setting can also introduce additional latency for certain operations, and timed events. It’s recommended not to touch this setting unless your CPU usage is unacceptable at idle workload. Some alternatives to this could be:

Reduce the number of worker threads (net-threads)
Reduce the number of disk (AIO) threads

Make sure accept threads are enabled

The relevant configurations for this are

records:
  exec_thread:
    autoconfig:
      enabled: 0
    limit: 2
  accept_threads: 1
  cache:
    threads_per_disk: 8

See Timeout Settings for more discussion on Traffic Server timeouts.

proxy.config.task_threads

Scope:

CONFIG

Type:

INT

Default:

2

yaml:

records:
  task_threads: 2

Specifies the number of task threads to run. These threads are used for various tasks that should be off-loaded from the normal network threads. You must have at least one task thread available.

proxy.config.allocator.thread_freelist_size

Scope:

CONFIG

Type:

INT

Default:

512

yaml:

records:
  allocator:
    thread_freelist_size: 512

Sets the maximum number of elements that can be contained in a ProxyAllocator (per-thread) before returning the objects to the global pool. If set to 0, there is no limit enforced.

proxy.config.allocator.thread_freelist_low_watermark

Scope:

CONFIG

Type:

INT

Default:

32

yaml:

records:
  allocator:
    thread_freelist_low_watermark: 32

Sets the minimum number of items a ProxyAllocator (per-thread) will guarantee to be holding at any one time.

proxy.config.allocator.hugepages

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  allocator:
    hugepages: 0

Enable (1) the use of huge pages on supported platforms. (Currently only Linux)

You must also enable hugepages at the OS level. In modern Linux kernels, this can be done by setting /proc/sys/vm/nr_hugepages and/or /proc/sys/vm/nr_overcommit_hugepages to a sufficiently large value. Check your system hugepage size (typically 2MB) using /proc/meminfo (look for Hugepagesize). The configured number of nr_hugepages plus nr_overcommit_hugepages times the hugepage size will be the amount of ram availabe for hugepages.

For more information on the implications of enabling huge pages, see Wikipedia <http://en.wikipedia.org/wiki/Page_%28computer_memory%29#Page_size_trade-off>_.

If hugepages are enabled, then the iobuffer allocator and cache directory entries will attempt to allocate hugepages for storage. If the number of hugepages is exhausted, the allocators will revert back to regular size pages.

proxy.config.dump_mem_info_frequency

Scope:

CONFIG

Type:

INT

Default:

0

Reloadable:

Yes

yaml:

records:
  dump_mem_info_frequency: 0

Enable <value>. When enabled makes Traffic Server dump IO Buffer memory information to traffic.out at <value> (intervals are in seconds). A zero value implies it is disabled

proxy.config.res_track_memory

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  res_track_memory: 0

When enabled makes Traffic Server track memory usage (allocations and releases). This information is dumped to traffic.out when the user sends a SIGUSR1 signal or periodically when proxy.config.dump_mem_info_frequency is enabled.

Value	Description
0	Memory tracking Disabled
1	Tracks IO Buffer Memory allocations and releases
2	Tracks IO Buffer Memory and OpenSSL Memory allocations and releases

proxy.config.system_clock

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  system_clock: 0

For advanced users only. This allows to specify the underlying system clock used by ATS. The default is CLOCK_REALTIME (0), but a higher performance option could be CLOCK_REALTIME_COARSE (5). See clock_gettime(2) for more details. On Linux, these definitions can be found in <linux/time.h>.

proxy.config.allocator.dontdump_iobuffers

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  allocator:
    dontdump_iobuffers: 1

Enable (1) the exclusion of IO buffers from core files when ATS crashes on supported platforms. (Currently only Linux). IO buffers are allocated with the MADV_DONTDUMP with madvise() on Linux platforms that support MADV_DONTDUMP. Enabled by default.

proxy.config.allocator.iobuf_chunk_sizes

Scope:

CONFIG

Type:

STRING

Default:

*NONE*

yaml:

records:
  allocator:
    iobuf_chunk_sizes: null

This configures the chunk sizes of each of the IO buffer allocators. The chunk size is the number of buffers allocated in a batch when the allocator’s freelist is exhausted. This must be specified as a space separated list of up to 15 numbers. If not specified or if any value specified is 0, the default value will be used.

The list of numbers will specify the chunk sizes in the following order:

128 256 512 1k 2k 4k 8k 16k 32k 64k 128k 256k 512k 1M 2M

The defaults for each allocator is:

128 128 128 128 128 128 32 32 32 32 32 32 32 32 32

Even though this is specified, the actual chunk size might be modified based on the system’s page size (or hugepage size if enabled).

You might want to adjust these values to reduce the overall number of allocations that ATS needs to make based on your configured RAM cache size. On a running system, you can send SIGUSR1 to the ATS process to have it log the allocator statistics and see how many of each buffer size have been allocated.

proxy.config.ssl.misc.io.max_buffer_index

Scope:

CONFIG

Type:

INT

Default:

8

yaml:

records:
  ssl:
    misc:
      io:
        max_buffer_index: 8

Configures the max IOBuffer Block index used for various SSL Operations such as Handshake or Protocol Probe. Default value is 8 which maps to a 32K buffer

proxy.config.hostdb.io.max_buffer_index

Scope:

CONFIG

Type:

INT

Default:

8

yaml:

records:
  hostdb:
    io:
      max_buffer_index: 8

Configures the max IOBuffer Block index used for storing HostDB records. Default value is 8 which maps to a 32K buffer

proxy.config.payload.io.max_buffer_index

Scope:

CONFIG

Type:

INT

Default:

8

yaml:

records:
  payload:
    io:
      max_buffer_index: 8

Configures the max IOBuffer Block index used for storing request payload buffer for a POST request. Default value is 8 which maps to a 32K buffer

proxy.config.msg.io.max_buffer_index

Scope:

CONFIG

Type:

INT

Default:

8

yaml:

records:
  msg:
    io:
      max_buffer_index: 8

Configures the max IOBuffer Block index used for storing miscellaneous transactional buffers such as error response body. Default value is 8 which maps to a 32K buffer

proxy.config.log.io.max_buffer_index

Scope:

CONFIG

Type:

INT

Default:

8

yaml:

records:
  log:
    io:
      max_buffer_index: 8

Configures the max IOBuffer Block index used for storing an access log entry. Default value is 8 which maps to a 32K buffer

proxy.config.http.enabled

Scope:

CONFIG

Type:

INT

Default:

1

yaml:

records:
  http:
    enabled: 1

Turn on or off support for HTTP proxying. This is rarely used, the one exception being if you run Traffic Server with a protocol plugin, and would like for it to not support HTTP requests at all.

proxy.config.http.allow_half_open

Scope:

CONFIG

Type:

INT

Default:

1

Reloadable:

Yes

Overridable:

Yes

yaml:

records:
  http:
    allow_half_open: 1

Turn on or off support for connection half open for client side. Default is on, so after client sends FIN, the connection is still there.

proxy.config.http.wait_for_cache

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  http:
    wait_for_cache: 0

Accepting inbound connections and starting the cache are independent operations in Traffic Server. This variable controls the relative timing of these operations and Traffic Server dependency on cache because if cache is required then inbound connection accepts should be deferred until the validity of the cache requirement is determined. Cache initialization failure will be logged in diags.log.

Value	Description
0	Decouple inbound connections and cache initialization. Connections will be accepted as soon as possible and Traffic Server will run regardless of the results of cache initialization.
1	Do not accept inbound connections until cache initialization has finished. Traffic Server will run regardless of the results of cache initialization.
2	Do not accept inbound connections until cache initialization has finished and been sufficiently successful that cache is enabled. This means at least one cache span is usable. If there are no spans in storage.config or none of the spans can be successfully parsed and initialized then Traffic Server will shut down.
3	Do not accept inbound connections until cache initialization has finished and been completely successful. This requires at least one cache span in storage.config and that every span specified is valid and successfully initialized. Any error will cause Traffic Server to shut down.

IO_URING

proxy.config.io_uring.entries

Scope:

CONFIG

Type:

INT

Default:

32

yaml:

records:
  io_uring:
    entries: 32

Specify the number of entries in each io_uring. There will be on io_uring instance per thread that uses io_uring for IO. This parameter is passed to io_uring_queue_init.

proxy.config.io_uring.sq_poll_ms

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  io_uring:
    sq_poll_ms: 0

If this value is >0 then use submit queue polling mode. The value will be used to specifiy the sq_thread_idle parameter to io_uring setup. More information about submit queue polling mode can be found here: https://unixism.net/loti/tutorial/sq_poll.html

proxy.config.io_uring.attach_wq

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  io_uring:
    attach_wq: 0

Set this to 1 if you want io_uring to re-use the same worker queue backend for each thread.

proxy.config.io_uring.wq_workers_bounded

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  io_uring:
    wq_workers_bounded: 0

proxy.config.io_uring.wq_workers_unbounded

Scope:

CONFIG

Type:

INT

Default:

0

yaml:

records:
  io_uring:
    wq_workers_unbounded: 0

These settings configured the number of threads for the io_uring worker queue backend. See the manpage for io_uring_register_iowq_max_workers for more information.

AIO

proxy.config.aio.mode

Scope:

CONFIG

Type:

STRING

Default:

auto

yaml:

records:
  aio:
    mode: auto

(Only if io_uring is enabled in the build) Normally, ATS will detect if io_uring can be used for async disk IO. Using this config item, the AIO mode can instead be specified. The value can be one of:

Value	Description
auto	Use the default detection logic
thread	Use the AIO thread pool for disk IO
io_uring	Use io_uring for disk IO

Note: If you force the backend to use io_uring, you might experience failures with some (older, pre 5.4) kernel versions