Feb 21, 2024: Announcing Linkerd 2.15 with support for VM workloads, native sidecars, and SPIFFE! Read more »

Proxy Configuration

Linkerd provides a set of annotations that can be used to override the data plane proxy’s configuration. This is useful for overriding the default configurations of auto-injected proxies.

The following is the list of supported annotations:

Annotation Description
config.alpha.linkerd.io/proxy-enable-native-sidecar Enable KEP-753 native sidecars. This is an experimental feature. It requires Kubernetes >= 1.29. If enabled, .proxy.waitBeforeExitSeconds should not be used.
config.alpha.linkerd.io/proxy-wait-before-exit-seconds The proxy sidecar will stay alive for at least the given period after receiving SIGTERM signal from Kubernetes but no longer than pod’s terminationGracePeriodSeconds. Defaults to 0
config.linkerd.io/access-log Enables HTTP access logging in the proxy. Accepted values are apache, to output the access log in the Appache Common Log Format, and json, to output the access log in JSON.
config.linkerd.io/admin-port Proxy port to serve metrics on
config.linkerd.io/close-wait-timeout Sets nf_conntrack_tcp_timeout_close_wait. Accepts a duration string, e.g. 1m or 3600s
config.linkerd.io/control-port Proxy port to use for control
config.linkerd.io/debug-image Linkerd debug container image name
config.linkerd.io/debug-image-pull-policy Docker image pull policy for debug image
config.linkerd.io/debug-image-version Linkerd debug container image version
config.linkerd.io/default-inbound-policy Proxy’s default inbound policy
config.linkerd.io/enable-debug-sidecar Inject a debug sidecar for data plane debugging
config.linkerd.io/enable-external-profiles Enable service profiles for non-Kubernetes services
config.linkerd.io/image-pull-policy Docker image pull policy
config.linkerd.io/inbound-port Proxy port to use for inbound traffic
config.linkerd.io/init-image Linkerd init container image name
config.linkerd.io/init-image-version Linkerd init container image version
config.linkerd.io/opaque-ports Ports that skip the proxy’s protocol detection mechanism and are proxied opaquely. Comma-separated list of values, where each value can be a port number or a range a-b.
config.linkerd.io/outbound-port Proxy port to use for outbound traffic
config.linkerd.io/pod-inbound-ports Comma-separated list of (non-proxy) container ports exposed by the pod spec. Useful when other mutating webhooks inject sidecar containers after the proxy injector has run
config.linkerd.io/proxy-await The application container will not start until the proxy is ready; accepted values are enabled and disabled
config.linkerd.io/proxy-cpu-limit Maximum amount of CPU units that the proxy sidecar can use
config.linkerd.io/proxy-cpu-request Amount of CPU units that the proxy sidecar requests
config.linkerd.io/proxy-disable-inbound-protocol-detect-timeout When set to true, disables the protocol detection timeout on the inbound side of the proxy by setting it to a very high value
config.linkerd.io/proxy-disable-outbound-protocol-detect-timeout When set to true, disables the protocol detection timeout on the outbound side of the proxy by setting it to a very high value
config.linkerd.io/proxy-ephemeral-storage-limit Used to override the limitEphemeralStorage config
config.linkerd.io/proxy-ephemeral-storage-request Used to override the requestEphemeralStorage config
config.linkerd.io/proxy-image Linkerd proxy container image name
config.linkerd.io/proxy-inbound-connect-timeout Inbound TCP connection timeout in the proxy
config.linkerd.io/proxy-inbound-discovery-cache-unused-timeout Maximum time allowed before an unused inbound discovery result is evicted from the cache. Defaults to 90s
config.linkerd.io/proxy-log-format Log format (plain or json) for the proxy
config.linkerd.io/proxy-log-level Log level for the proxy
config.linkerd.io/proxy-memory-limit Maximum amount of Memory that the proxy sidecar can use
config.linkerd.io/proxy-memory-request Amount of Memory that the proxy sidecar requests
config.linkerd.io/proxy-outbound-connect-timeout Used to configure the outbound TCP connection timeout in the proxy
config.linkerd.io/proxy-outbound-discovery-cache-unused-timeout Maximum time allowed before an unused outbound discovery result is evicted from the cache. Defaults to 5s
config.linkerd.io/proxy-uid Run the proxy under this user ID
config.linkerd.io/proxy-version Tag to be used for the Linkerd proxy images
config.linkerd.io/shutdown-grace-period Grace period for graceful proxy shutdowns. If this timeout elapses before all open connections have completed, the proxy will terminate forcefully, closing any remaining connections.
config.linkerd.io/skip-inbound-ports Ports that should skip the proxy and send directly to the application. Comma-separated list of values, where each value can be a port number or a range a-b.
config.linkerd.io/skip-outbound-ports Outbound ports that should skip the proxy. Comma-separated list of values, where each value can be a port number or a range a-b.
config.linkerd.io/skip-subnets Comma-separated list of subnets in valid CIDR format that should be skipped by the proxy
linkerd.io/inject Controls whether or not a pod should be injected; accepted values are enabled, disabled and ingress

For example, to update an auto-injected proxy’s CPU and memory resources, we insert the appropriate annotations into the spec.template.metadata.annotations of the owner’s pod spec, using kubectl edit like this:

        config.linkerd.io/proxy-cpu-limit: "1"
        config.linkerd.io/proxy-cpu-request: "0.2"
        config.linkerd.io/proxy-memory-limit: 2Gi
        config.linkerd.io/proxy-memory-request: 128Mi

See here for details on tuning the proxy’s resource usage.

For proxies injected using the linkerd inject command, configuration can be overridden using the command-line flags.

Ingress Mode

Proxy ingress mode is a mode of operation designed to help Linkerd integrate with certain ingress controllers. Ingress mode is necessary if the ingress itself cannot be otherwise configured to use the Service port/ip as the destination.

When an individual Linkerd proxy is set to ingress mode, it will route requests based on their :authority, Host, or l5d-dst-override headers instead of their original destination. This will inform Linkerd to override the endpoint selection of the ingress container and to perform its own endpoint selection, enabling features such as per-route metrics and traffic splitting.

The proxy can be made to run in ingress mode by using the linkerd.io/inject: ingress annotation rather than the default linkerd.io/inject: enabled annotation. This can also be done with the --ingress flag in the inject CLI command:

kubectl get deployment <ingress-controller> -n <ingress-namespace> -o yaml | linkerd inject --ingress - | kubectl apply -f -