MeshFaultInjection

With the MeshFaultInjection policy you can easily test your microservices against resiliency.

This policy uses a new policy matching algorithm. Do not combine with FaultInjection.

targetRef support matrix

targetRefAllowed kinds
targetRef.kindMesh, MeshSubset
from[].targetRef.kindMesh, MeshSubset, MeshServiceSubset
targetRefAllowed kinds
targetRef.kindMesh, MeshGateway, MeshGateway with listener tags
to[].targetRef.kindMesh

MeshFaultInjection isn’t supported on delegated gateways.

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

Configuration

MeshFaultInjection allows configuring a list of HTTP faults. They execute in the same order as they were defined.

  1. default:
  2.   http:
  3.     - abort:
  4.         httpStatus: 500
  5.         percentage: "2.5"
  6.       delay:
  7.         value: 5s
  8.         percentage: 5
  9.       responseBandwidth:
  10.         limit: "50Mbps"
  11.         percentage: 50
  12.     - abort:
  13.         httpStatus: 500
  14.         percentage: 10
  15.     - delay:
  16.         value: 5s
  17.         percentage: 5

It’s worth mentioning that percentage of the next filter depends on the percentage of previous ones.

  1. http:
  2.   - abort:
  3.       httpStatus: 500
  4.       percentage: 70
  5.   - abort:
  6.       httpStatus: 503
  7.       percentage: 50

That means that for 70% of requests, it returns 500 and for 50% of the 30% that passed it returns 503.

Abort

Abort defines a configuration of not delivering requests to destination service and replacing the responses from destination data plane by predefined status code.

  • httpStatus - HTTP status code which will be returned to source side, has to be in [100 - 599] range
  • percentage - a percentage of requests on which abort will be injected, has to be in [0.0 - 100.0] range. If the value is a double number, put it in quotes.

Delay

Delay defines a configuration of delaying a response from a destination.

  • value - the duration during which the response will be delayed
  • percentage - a percentage of requests on which abort will be injected, has to be in [0.0 - 100.0] range. If the value is a double number, put it in quotes.

ResponseBandwidth limit

ResponseBandwidth defines a configuration to limit the speed of responding to requests.

  • limit - represented by value measure in Gbps, Mbps, kbps, or bps, for example 10kbps
  • percentage - a percentage of requests on which abort will be injected, has to be in [0.0 - 100.0] range. If the value is a double number, put it in quotes.

Examples

Service backend returns 500 for 50% of requests from frontend service

  1. apiVersion: kuma.io/v1alpha1
  2. kind: MeshFaultInjection
  3. metadata:
  4.   name: default-fault-injection
  5.   namespace: kuma-demo
  6.   labels:
  7.     kuma.io/mesh: default
  8. spec:
  9.   targetRef:
  10.     kind: MeshSubset
  11.     tags:
  12.       app: backend
  13.   from:
  14.   - targetRef:
  15.       kind: MeshSubset
  16.       tags:
  17.         kuma.io/service: frontend
  18.     default:
  19.       http:
  20.       - abort:
  21.           httpStatus: 500
  22.           percentage: 50
  1. type: MeshFaultInjection
  2. mesh: default
  3. name: default-fault-injection
  4. spec:
  5.   targetRef:
  6.     kind: MeshSubset
  7.     tags:
  8.       app: backend
  9.   from:
  10.   - targetRef:
  11.       kind: MeshSubset
  12.       tags:
  13.         kuma.io/service: frontend
  14.     default:
  15.       http:
  16.       - abort:
  17.           httpStatus: 500
  18.           percentage: 50

50.5% of requests to service backend from any service is going to be delayed by 5 seconds

  1. apiVersion: kuma.io/v1alpha1
  2. kind: MeshFaultInjection
  3. metadata:
  4.   name: default-fault-injection
  5.   namespace: kuma-demo
  6.   labels:
  7.     kuma.io/mesh: default
  8. spec:
  9.   targetRef:
  10.     kind: MeshSubset
  11.     tags:
  12.       app: backend
  13.   from:
  14.   - targetRef:
  15.       kind: Mesh
  16.       name: default
  17.     default:
  18.       http:
  19.       - delay:
  20.           percentage: '50.5'
  21.           value: 5s
  1. type: MeshFaultInjection
  2. mesh: default
  3. name: default-fault-injection
  4. spec:
  5.   targetRef:
  6.     kind: MeshSubset
  7.     tags:
  8.       app: backend
  9.   from:
  10.   - targetRef:
  11.       kind: Mesh
  12.       name: default
  13.     default:
  14.       http:
  15.       - delay:
  16.           percentage: '50.5'
  17.           value: 5s

Backend service with a list of faults that are applied for frontend service

  1. apiVersion: kuma.io/v1alpha1
  2. kind: MeshFaultInjection
  3. metadata:
  4.   name: default-fault-injection
  5.   namespace: kuma-demo
  6.   labels:
  7.     kuma.io/mesh: default
  8. spec:
  9.   targetRef:
  10.     kind: MeshSubset
  11.     tags:
  12.       app: backend
  13.   from:
  14.   - targetRef:
  15.       kind: MeshSubset
  16.       tags:
  17.         kuma.io/service: frontend
  18.     default:
  19.       http:
  20.       - abort:
  21.           httpStatus: 500
  22.           percentage: '2.5'
  23.       - abort:
  24.           httpStatus: 500
  25.           percentage: 10
  26.       - delay:
  27.           value: 5s
  28.           percentage: 5
  1. type: MeshFaultInjection
  2. mesh: default
  3. name: default-fault-injection
  4. spec:
  5.   targetRef:
  6.     kind: MeshSubset
  7.     tags:
  8.       app: backend
  9.   from:
  10.   - targetRef:
  11.       kind: MeshSubset
  12.       tags:
  13.         kuma.io/service: frontend
  14.     default:
  15.       http:
  16.       - abort:
  17.           httpStatus: 500
  18.           percentage: '2.5'
  19.       - abort:
  20.           httpStatus: 500
  21.           percentage: 10
  22.       - delay:
  23.           value: 5s
  24.           percentage: 5

All policy options