# Bookstore — Part 11 ch.07 "Chaos engineering": a Chaos Mesh WORKFLOW that
# runs the three experiments in sequence as a repeatable GAME DAY / CI gate —
# the discipline (hypothesis → inject → observe → learn), automated.
#
# WHY A WORKFLOW (not three ad-hoc applies)
#   The chapter's point: chaos is a DISCIPLINE, not random breakage. A
#   Workflow encodes the run as code — ordered, time-bounded, abortable,
#   reproducible — so it can be a CI/CD GATE (Part 07): promote only if the
#   Bookstore survived pod-kill + network latency + CPU stress with its SLO
#   intact (raw-manifests/81-prometheusrule.yaml). Sequential (not parallel)
#   keeps the blast radius one-experiment-at-a-time; each child has its own
#   bounded duration and is the SAME spec as 10-/20-/30-.
#
# ABORT / SAFETY: deleting this Workflow object stops the run and reverts any
# in-flight child experiment (Chaos Mesh tears down injected rules on
# stop/expiry). In production you ALSO wire an abort condition (halt the game
# day if the real error budget burns — Part 06 ch.05). Nothing here mutates
# the Bookstore Pods; all children are restricted-safe (as 10-/20-/30-).
#
# !!! CRD-INTRINSIC DRY-RUN (identical precedent to 18-/51-/70-/83-/argocd) !!!
#   `Workflow` is a Chaos Mesh CRD (chaos-mesh.org/v1alpha1). WITHOUT Chaos
#   Mesh installed a client dry-run prints:
#     no matches for kind "Workflow" in version "chaos-mesh.org/v1alpha1"
#   EXPECTED, schema-correct — Chaos Mesh CRDs/controller must be installed
#   first. Whole-dir dry-run prints this for CRD-backed files only and
#   continues. Schema verified against Chaos Mesh chaos-mesh.org/v1alpha1
#   Workflow (serial entry; embedded *Chaos templates).
#
# Requires: Chaos Mesh installed (own ns); the Bookstore running with PDB
#   (84-), HPA (82-), metrics/SLO (80-/81-) so each step is observable.
# Apply (the whole game day, then watch it in the Chaos Dashboard):
#   kubectl apply -f examples/bookstore/chaos/40-workflow-gameday.yaml
#   kubectl get workflow,workflownode -n bookstore
#   kubectl delete -f examples/bookstore/chaos/40-workflow-gameday.yaml   # abort
apiVersion: chaos-mesh.org/v1alpha1
kind: Workflow
metadata:
  name: bookstore-gameday
  namespace: bookstore
  labels:
    app.kubernetes.io/part-of: bookstore
spec:
  entry: gameday-sequence
  templates:
    # Run the three experiments IN ORDER (blast radius = one at a time).
    - name: gameday-sequence
      templateType: Serial
      deadline: 600s                 # hard ceiling on the whole game day
      children:
        - kill-one-catalog
        - catalog-postgres-latency
        - catalog-cpu-stress

    # --- step 1: pod-kill (same spec as 10-) ------------------------------
    - name: kill-one-catalog
      templateType: PodChaos
      deadline: 60s
      podChaos:
        action: pod-kill
        mode: one
        selector:
          namespaces: [bookstore]
          labelSelectors: { app: catalog }

    # --- step 2: network latency (same spec as 20-) -----------------------
    - name: catalog-postgres-latency
      templateType: NetworkChaos
      deadline: 90s
      networkChaos:
        action: delay
        mode: all
        selector:
          namespaces: [bookstore]
          labelSelectors: { app: catalog }
        direction: to
        target:
          mode: all
          selector:
            namespaces: [bookstore]
            labelSelectors: { app: postgres }
        delay: { latency: "200ms", jitter: "50ms", correlation: "50" }
        duration: "60s"

    # --- step 3: CPU stress (same spec as 30-) ----------------------------
    - name: catalog-cpu-stress
      templateType: StressChaos
      deadline: 90s
      stressChaos:
        mode: one
        selector:
          namespaces: [bookstore]
          labelSelectors: { app: catalog }
        stressors:
          cpu: { workers: 2, load: 100 }
        duration: "60s"