Kuma Star

Careful!

You are browsing documentation for the next version of Kuma. Use this version at your own risk.

Mesh HTTP Route

This policy uses new policy matching algorithm. Do not combine with TrafficRoute except for the default route-all route, which should be kept..

The MeshHTTPRoute policy allows altering and redirecting HTTP requests depending on where the request coming from and where it’s going to.

TargetRef support matrix

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

If you don’t understand this table you should read matching docs.

Configuration

Unlike others outbound policies MeshHTTPRoute doesn’t contain default directly in the to array. The default section is nested inside rules, so the policy structure looks like this:

spec:
  targetRef: # top-level targetRef selects a group of proxies to configure
    kind: Mesh|MeshSubset|MeshService|MeshServiceSubset
  to:
    - targetRef: # targetRef selects a destination (outbound listener)
        kind: MeshService
        name: backend
      rules:
        - matches: [...] # various ways to match an HTTP request (path, method, query)
          default: # configuration applied for the matched HTTP request
            filters: [...]
            backendRefs: [...]

Remember to tag your Service ports with appProtocol: http to use them in a MeshHTTPRoute!

Matches

  • path - (optional) - HTTP path to match the request on
    • type - one of Exact, PathPrefix , RegularExpression
    • value - actual value that’s going to be matched depending on the type
  • method - (optional) - HTTP2 method, available values are CONNECT, DELETE, GET, HEAD, OPTIONS, PATCH, POST, PUT, TRACE
  • queryParams - (optional) - list of HTTP URL query parameters. Multiple matches are ANDed together such that all listed matches must succeed
    • type - one of Exact or RegularExpression
    • name - name of the query parameter
    • value - actual value that’s going to be matched depending on the type

Default conf

  • filters - (optional) - a list of modifications applied to the matched request
    • type - available values are RequestHeaderModifier, ResponseHeaderModifier, RequestRedirect, URLRewrite.
    • requestHeaderModifier - HeaderModifier, must be set if the type is RequestHeaderModifier.
    • responseHeaderModifier - HeaderModifier, must be set if the type is ResponseHeaderModifier.
    • requestRedirect - must be set if the type is RequestRedirect
      • scheme - one of http or http2
      • hostname - is the fully qualified domain name of a network host. This matches the RFC 1123 definition of a hostname with 1 notable exception that numeric IP addresses are not allowed.
      • port - is the port to be used in the value of the Location header in the response. When empty, port (if specified) of the request is used.
      • statusCode - is the HTTP status code to be used in response. Available values are 301, 302, 303, 307, 308.
    • urlRewrite - must be set if the type is URLRewrite
      • hostname - (optional) - is the fully qualified domain name of a network host. This matches the RFC 1123 definition of a hostname with 1 notable exception that numeric IP addresses are not allowed.
      • path - (optional)
        • type - one of ReplaceFullPath, ReplacePrefixMatch
        • replaceFullPath - must be set if the type is ReplaceFullPath
        • replacePrefixMatch - must be set if the type is ReplacePrefixMatch
    • requestMirror - must be set if the type is RequestMirror
      • percentage - percentage of requests to mirror. If not specified, all requests to the target cluster will be mirrored.
      • backendRef - BackendRef, destination to mirror request to
  • backendRefs - BackendRef (optional), list of destinations to redirect requests to

HeaderModifier

  • set - (optional) - list of headers to set. Overrides value if the header exists.
    • name - header’s name
    • value - header’s value
  • add - (optional) - list of headers to add. Appends value if the header exists.
    • name - header’s name
    • value - header’s value
  • remove - (optional) - list of headers’ names to remove

BackendRef

  • kind - one of MeshService, MeshServiceSubset
  • name - service name
  • tags - service tags, must be specified if the kind is MeshServiceSubset
  • weight - when a request matches the route, the choice of an upstream cluster is determined by its weight. Total weight is a sum of all weights in backendRefs list.

Interactions with MeshTCPRoute

MeshHTTPRoute takes priority over MeshTCPRoute when both are defined for the same service, and the matching MeshTCPRoute is ignored.

Examples

Traffic split

We can use MeshHTTPRoute to split an HTTP traffic between services with different tags implementing A/B testing or canary deployments.

Here is an example of a MeshHTTPRoute that splits the traffic from frontend_kuma-demo_svc_8080 to backend_kuma-demo_svc_3001 between versions, but only on endpoints starting with /api. All other endpoints will go to version: 1.0.

apiVersion: kuma.io/v1alpha1
kind: MeshHTTPRoute
metadata:
  name: http-route-1
  namespace: kuma-system
  labels:
    kuma.io/mesh: default
spec:
  targetRef:
    kind: MeshService
    name: frontend_kuma-demo_svc_8080
  to:
    - targetRef:
        kind: MeshService
        name: backend_kuma-demo_svc_3001
      rules:
        - matches:
            - path:
                type: PathPrefix
                value: /
          default:
            backendRefs:
              - kind: MeshServiceSubset
                name: backend_kuma-demo_svc_3001
                tags:
                  version: "v0"
                weight: 90
              - kind: MeshServiceSubset
                name: backend_kuma-demo_svc_3001
                tags:
                  version: "v1"
                weight: 10

Traffic modifications

We can use MeshHTTPRoute to modify outgoing requests, by setting new path or changing request and response headers.

Here is an example of a MeshHTTPRoute that adds x-custom-header with value xyz when frontend_kuma-demo_svc_8080 tries to consume backend_kuma-demo_svc_3001.

apiVersion: kuma.io/v1alpha1
kind: MeshHTTPRoute
metadata:
  name: http-route-1
  namespace: kuma-system
  labels:
    kuma.io/mesh: default
spec:
  targetRef:
    kind: MeshService
    name: frontend_kuma-demo_svc_8080
  to:
    - targetRef:
        kind: MeshService
        name: backend_kuma-demo_svc_3001
      rules:
        - matches:
            - path:
                type: Exact
                value: /
          default:
            filters:
              - type: RequestHeaderModifier
                requestHeaderModifier:
                  set:
                    - name: x-custom-header
                      value: xyz

Traffic mirror

MeshHTTPRoute can mirror a fraction of requests to another service. This can be useful when testing a new version of the app with the production payload without interrupting real users.

apiVersion: kuma.io/v1alpha1
kind: MeshHTTPRoute
metadata:
  name: http-route-1
  namespace: kuma-system
  labels:
    kuma.io/mesh: default
spec:
  targetRef:
    kind: MeshService
    name: frontend_kuma-demo_svc_8080
  to:
    - targetRef:
        kind: MeshService
        name: backend_kuma-demo_svc_3001
      rules:
        - matches:
            - headers:
                - type: Exact
                  name: mirror-this-request
                  value: "true"
          default:
            filters:
              - type: RequestMirror
                requestMirror:
                  percentage: 30
                  backendRef:
                    kind: MeshServiceSubset
                    name: backend_kuma-demo_svc_3001
                    tags:
                      version: v1_experimental
            backendRefs:
              - kind: MeshServiceSubset
                name: backend_kuma-demo_svc_3001
                tags:
                  version: v0

Merging

When several MeshHTTPRoute policies target the same data plane proxy they’re merged. Similar to the new policies the merging order is determined by the top level targetRef. The difference is in spec.to[].rules. Kuma treats rules as a key-value map where matches is a key and default is a value. For example MeshHTTPRoute policies:

# MeshHTTPRoute-1
rules:
  - matches: # key-1
      - path:
          type: Exact
          name: /orders
        method: GET
    default: CONF_1 # value
  - matches: # key-2
      - path:
          type: Exact
          name: /payments
        method: POST
    default: CONF_2 # value
---
# MeshHTTPRoute-2
rules:
  - matches: # key-3
      - path:
          type: Exact
          name: /orders
        method: GET
    default: CONF_3 # value
  - matches: # key-4
      - path:
          type: Exact
          name: /payments
        method: POST
    default: CONF_4 # value

merged in the following list of rules:

rules:
  - matches:
      - path:
          type: Exact
          name: /orders
        method: GET
    default: merge(CONF_1, CONF_3) # because 'key-1' == 'key-3'
  - matches:
      - path:
          type: Exact
          name: /payments
        method: POST
    default: merge(CONF_2, CONF_4) # because 'key-2' == 'key-4'

All policy options