Kuma Star

MeshRetry

This policy uses new policy matching algorithm. Do not combine with Retry.

This policy enables Kuma to know how to behave if there is a failed scenario (i.e. HTTP request) which could be retried.

TargetRef support matrix

targetRef Allowed kinds
targetRef.kind Mesh, MeshSubset, MeshService, MeshServiceSubset
to[].targetRef.kind Mesh, MeshService

To learn more about the information in this table, see the matching docs.

Configuration

The policy let you configure retry behaviour for HTTP, GRPC and TCP protocols. The protocol is selected by picking the most specific protocol.

Each protocol has a separate section under default in the policy YAML. Some sections are common between protocols or have similar meaning.

Retry on

The field retryOn is a list of conditions which will cause a retry.

For HTTP these are related to the response status code or method (“5xx”, “429”, “HttpMethodGet”). For gRPC these are status codes in response headers (“canceled”, “deadline-exceeded”, etc.). There is no equivalent for TCP.

One or more conditions can be specified, for example:

retryOn:
  - "429"
  - "503"

means that it the policy will retry on a status code 429 or 503.

Full list of available HTTP conditions:

retryOn:
  - 5XX
  - GatewayError
  - Reset
  - Retriable4xx
  - ConnectFailure
  - EnvoyRatelimited
  - RefusedStream
  - Http3PostConnectFailure
  - HttpMethodConnect
  - HttpMethodDelete
  - HttpMethodGet
  - HttpMethodHead
  - HttpMethodOptions
  - HttpMethodPatch
  - HttpMethodPost
  - HttpMethodPut
  - HttpMethodTrace
  - "429" # any HTTP status code
  - "503"

Full list of available gRPC conditions:

retryOn:
  - Canceled
  - DeadlineExceeded
  - Internal
  - ResourceExhausted
  - Unavailable

Backoff

This parameter is applicable to both HTTP and GRPC.

It consists of BaseInterval (the amount of time between retries) and MaxInterval (the maximal amount of time taken between retries).

We use a fully jittered exponential back-off algorithm for retries. Given a base interval B and retry number N, the back-off for the retry is in the range [0, (2N - 1) × B).

For example, given a 25ms interval, the first retry will be delayed randomly by 0-24ms, the 2nd by 0-74ms, the 3rd by 0-174ms, and so on.

The interval is capped at a MaxInterval, which defaults to 10 times the BaseInterval.

Rate limited backoff

This parameter is applicable to both HTTP and GRPC.

MeshRetry can be configured in such a way that when the upstream server rate limits the request and responds with a header like retry-after or x-ratelimit-reset it uses the value from the header to determine when to send the retry request instead of the backoff algorithm.

Example

Given this configuration:

retryOn:
  - "503"
rateLimitedBackOff:
  resetHeaders:
    - name: retry-after
      format: Seconds
    - name: x-ratelimit-reset
      format: UnixTimestamp

and an HTTP response:

HTTP/1.1 503 Service Unavailable
retry-after: 15

The retry request will be issued after 15 seconds.

If the response is as follows:

HTTP/1.1 503 Service Unavailable
x-ratelimit-reset: 1706096119

The request will be retried at Wed Jan 24 2024 11:35:19 GMT+0000.

If the response does not contain retry-after or x-ratelimit-reset header (with valid integer value) then the amount of time to wait before issuing a request is determined by backoff algorithm.

Examples

HTTP web to backend on 5xx

apiVersion: kuma.io/v1alpha1
kind: MeshRetry
metadata:
  name: web-to-backend-retry-http
  namespace: kuma-system
  labels:
    kuma.io/mesh: default # optional, defaults to `default` if unset
spec:
  targetRef:
    kind: MeshService
    name: web
  to:
    - targetRef:
        kind: MeshService
        name: backend
      default:
        http:
          numRetries: 10
          backOff:
            baseInterval: 15s
            maxInterval: 20m
          retryOn:
            - "5xx"

Apply the configuration with kubectl apply -f [..].

gRPC web to backend on DeadlineExceeded

apiVersion: kuma.io/v1alpha1
kind: MeshRetry
metadata:
  name: web-to-backend-retry-grpc
  namespace: kuma-system
  labels:
    kuma.io/mesh: default # optional, defaults to `default` if unset
spec:
  targetRef:
    kind: MeshService
    name: web
  to:
    - targetRef:
        kind: MeshService
        name: backend
      default:
        grpc:
          numRetries: 5
          backOff:
            baseInterval: 5s
            maxInterval: 1m
          retryOn:
            - "DeadlineExceeded"

Apply the configuration with kubectl apply -f [..].

TCP web to backend

apiVersion: kuma.io/v1alpha1
kind: MeshRetry
metadata:
  name: web-to-backend-retry-tcp
  namespace: kuma-system
  labels:
    kuma.io/mesh: default # optional, defaults to `default` if unset
spec:
  targetRef:
    kind: MeshService
    name: web
  to:
    - targetRef:
        kind: MeshService
        name: backend
      default:
        tcp:
          maxConnectAttempt: 5

Apply the configuration with kubectl apply -f [..].

All policy options