data-science-pipelines Custom Experiments

This page provides templates for writing custom chaos experiments targeting data-science-pipelines.

data-science-pipelines-operator

apiVersion: chaos.operatorchaos.io/v1alpha1
kind: ChaosExperiment
metadata:
  name: data-science-pipelines-operator-custom
spec:
  target:
    operator: data-science-pipelines
    component: data-science-pipelines-operator
  steadyState:
    checks:
      - type: conditionTrue
        apiVersion: apps/v1
        kind: Deployment
        name: data-science-pipelines-operator-controller-manager
        namespace: opendatahub
        conditionType: Available
    timeout: "60s"
  injection:
    type: PodKill  # Change to desired injection type
    parameters:
      labelSelector: app=data-science-pipelines-operator
    ttl: "300s"
  hypothesis:
    description: >-
      Describe the expected behavior after fault injection.
    recoveryTimeout: 120s

ds-pipelines-webhook

apiVersion: chaos.operatorchaos.io/v1alpha1
kind: ChaosExperiment
metadata:
  name: ds-pipelines-webhook-custom
spec:
  target:
    operator: data-science-pipelines
    component: ds-pipelines-webhook
  steadyState:
    checks:
      - type: conditionTrue
        apiVersion: apps/v1
        kind: Deployment
        name: ds-pipelines-webhook
        namespace: opendatahub
        conditionType: Available
    timeout: "60s"
  injection:
    type: PodKill  # Change to desired injection type
    parameters:
      labelSelector: app=ds-pipelines-webhook
    ttl: "300s"
  hypothesis:
    description: >-
      Describe the expected behavior after fault injection.
    recoveryTimeout: 120s

Running Custom Experiments

1. Save your experiment YAML to a file
2. Run: chaos-cli run --experiment <file>
3. Check results: chaos-cli results --latest