# Bookstore — Part 11 ch.07 "Chaos engineering": EXPERIMENT 2 — inject 200ms
# latency on catalog's traffic for 60s and observe the catalog SLO. Tests
# whether the system degrades GRACEFULLY (slower, not down) when the
# catalog→postgres path is impaired.
#
# THE HYPOTHESIS THIS TESTS
#   "Adding 200ms±50ms network latency to catalog Pods for 60s degrades
#   catalog p95 latency but does NOT push the 5xx error ratio over the SLO
#   (raw-manifests/81-prometheusrule.yaml: CatalogHighErrorRate fires at >5%
#   5xx for 10m; CatalogHighLatencyP95 at the p95 threshold). I.e. timeouts
#   and pooling absorb a slow dependency rather than cascading to failure."
#   Latency is the most common real-world degradation — far more than a clean
#   Pod kill — so this is the higher-value experiment.
#
# SELECTOR matches the REAL catalog pod label (app=catalog, ns bookstore;
# verified against raw-manifests/10-catalog-deploy.yaml). `direction: to` +
# the postgres-side selector scope the impairment to the catalog→postgres
# EDGE specifically (postgres app label = `postgres`, verified against
# raw-manifests/20-postgres-statefulset.yaml / 40-services.yaml), not all of
# catalog's traffic — a tighter blast radius.
#
# BOUNDED & REVERSIBLE: `duration: 60s` auto-reverts; Chaos Mesh removes the
# tc/netem rule when the experiment ends or the object is deleted. The
# injection is via the Chaos Mesh DAEMON (its own privileged DaemonSet in the
# chaos-mesh namespace — NOT bookstore); the catalog Pods themselves are NOT
# mutated and stay PSA-restricted-compliant. `kubectl delete -f` aborts early.
#
# !!! CRD-INTRINSIC DRY-RUN (identical precedent to 18-/51-/70-/83-/argocd) !!!
#   `NetworkChaos` is a Chaos Mesh CRD (chaos-mesh.org/v1alpha1). WITHOUT
#   Chaos Mesh installed a client dry-run prints:
#     no matches for kind "NetworkChaos" 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
#   NetworkChaos (action: delay).
#
# Requires: Chaos Mesh installed (own ns); the Bookstore running (catalog +
#   postgres + 80-/81- so the SLO series exist to observe).
# Apply:
#   kubectl apply -f examples/bookstore/chaos/20-networkchaos-latency-catalog-postgres.yaml
#   # observe: catalog p95 rises in the metrics; CatalogHighErrorRate should
#   #          NOT fire if timeouts/pooling are correct (the hypothesis).
#   kubectl delete -f examples/bookstore/chaos/20-networkchaos-latency-catalog-postgres.yaml
apiVersion: chaos-mesh.org/v1alpha1
kind: NetworkChaos
metadata:
  name: catalog-to-postgres-latency
  namespace: bookstore
  labels:
    app.kubernetes.io/part-of: bookstore
spec:
  action: delay
  mode: all                       # all catalog Pods (the edge under test)
  selector:
    namespaces:
      - bookstore
    labelSelectors:
      app: catalog                # REAL catalog pod label (10-)
  direction: to                   # impair catalog → <target> traffic…
  target:                         # …specifically the postgres edge
    mode: all
    selector:
      namespaces:
        - bookstore
      labelSelectors:
        app: postgres             # REAL postgres pod label (20-)
  delay:
    latency: "200ms"
    jitter: "50ms"
    correlation: "50"
  duration: "60s"                 # BOUNDED — auto-reverts after 60s