App shop-demo con guasti

Un componente sano di riferimento e sei componenti, ciascuno rotto in modo diverso. La tabella è il “risultato atteso” contro cui valutare i tool AI.

Componente	Guasto	Sintomo	Segnale principale
catalog	nessuno	Running, Ready	baseline
frontend	tag immagine inesistente	ImagePullBackOff	evento Failed to pull image
api	memory limit sotto il consumo	OOMKilled → CrashLoopBackOff	lastState.terminated.reason: OOMKilled
worker	ConfigMap mancante	CreateContainerConfigError	evento configmap "worker-config" not found
checkout	readiness probe su porta sbagliata	Running, 0/1 Ready	evento Readiness probe failed
checkout-svc	selector errato	Service senza endpoint	endpoints vuoti
orders-db	PVC su StorageClass inesistente	PVC e pod Pending	evento storageclass "fast-ssd" not found

Manifest

mkdir -p ~/Downloads/K8S-AI/lab && cd ~/Downloads/K8S-AI/lab

cat > shop-demo.yaml <<'EOF2'
apiVersion: v1
kind: Namespace
metadata:
  name: shop-demo
  labels:
    purpose: ai-debug-lab
---
# baseline: healthy
apiVersion: apps/v1
kind: Deployment
metadata:
  name: catalog
  namespace: shop-demo
spec:
  replicas: 1
  selector:
    matchLabels: { app: catalog }
  template:
    metadata:
      labels: { app: catalog }
    spec:
      containers:
        - name: podinfo
          image: ghcr.io/stefanprodan/podinfo:6.7.1
          ports: [{ containerPort: 9898, name: http }]
          readinessProbe:
            httpGet: { path: /readyz, port: http }
          resources:
            requests: { cpu: 10m, memory: 32Mi }
            limits: { memory: 64Mi }
---
apiVersion: v1
kind: Service
metadata:
  name: catalog
  namespace: shop-demo
spec:
  selector: { app: catalog }
  ports: [{ port: 80, targetPort: http }]
---
# fault 1: image tag does not exist
apiVersion: apps/v1
kind: Deployment
metadata:
  name: frontend
  namespace: shop-demo
spec:
  replicas: 1
  selector:
    matchLabels: { app: frontend }
  template:
    metadata:
      labels: { app: frontend }
    spec:
      containers:
        - name: nginx
          image: nginx:1.27-doesnotexist
          resources:
            requests: { cpu: 10m, memory: 16Mi }
---
# fault 2: memory limit below real usage -> OOMKilled
apiVersion: apps/v1
kind: Deployment
metadata:
  name: api
  namespace: shop-demo
spec:
  replicas: 1
  selector:
    matchLabels: { app: api }
  template:
    metadata:
      labels: { app: api }
    spec:
      containers:
        - name: stress
          image: polinux/stress
          command: ["stress", "--vm", "1", "--vm-bytes", "200M", "--vm-hang", "1"]
          resources:
            requests: { cpu: 10m, memory: 32Mi }
            limits: { memory: 64Mi }
---
# fault 3: references a ConfigMap that does not exist
apiVersion: apps/v1
kind: Deployment
metadata:
  name: worker
  namespace: shop-demo
spec:
  replicas: 1
  selector:
    matchLabels: { app: worker }
  template:
    metadata:
      labels: { app: worker }
    spec:
      containers:
        - name: worker
          image: busybox:1.36
          command: ["sh", "-c", "echo queue=$QUEUE_URL; sleep 3600"]
          env:
            - name: QUEUE_URL
              valueFrom:
                configMapKeyRef: { name: worker-config, key: queue_url }
          resources:
            requests: { cpu: 5m, memory: 8Mi }
---
# fault 4: readiness probe on the wrong port
apiVersion: apps/v1
kind: Deployment
metadata:
  name: checkout
  namespace: shop-demo
spec:
  replicas: 1
  selector:
    matchLabels: { app: checkout }
  template:
    metadata:
      labels: { app: checkout }
    spec:
      containers:
        - name: podinfo
          image: ghcr.io/stefanprodan/podinfo:6.7.1
          ports: [{ containerPort: 9898, name: http }]
          readinessProbe:
            httpGet: { path: /readyz, port: 8080 }
            periodSeconds: 5
          resources:
            requests: { cpu: 10m, memory: 32Mi }
            limits: { memory: 64Mi }
---
# fault 5: selector does not match any pod
apiVersion: v1
kind: Service
metadata:
  name: checkout-svc
  namespace: shop-demo
spec:
  selector: { app: check-out }
  ports: [{ port: 80, targetPort: 9898 }]
---
# fault 6: PVC on a StorageClass that does not exist
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: orders-db-data
  namespace: shop-demo
spec:
  accessModes: [ReadWriteOnce]
  storageClassName: fast-ssd
  resources:
    requests: { storage: 1Gi }
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: orders-db
  namespace: shop-demo
spec:
  replicas: 1
  selector:
    matchLabels: { app: orders-db }
  template:
    metadata:
      labels: { app: orders-db }
    spec:
      containers:
        - name: redis
          image: redis:7
          volumeMounts: [{ name: data, mountPath: /data }]
          resources:
            requests: { cpu: 10m, memory: 32Mi }
      volumes:
        - name: data
          persistentVolumeClaim: { claimName: orders-db-data }
EOF2

kubectl apply -f shop-demo.yaml

Verifica dello stato rotto

kubectl -n shop-demo get pods,svc,endpoints,pvc

kubectl -n shop-demo get events --sort-by=.lastTimestamp | tail -25

kubectl -n shop-demo get pods -o custom-columns=NAME:.metadata.name,READY:.status.containerStatuses[0].ready,RESTARTS:.status.containerStatuses[0].restartCount,WAITING:.status.containerStatuses[0].state.waiting.reason,LAST:.status.containerStatuses[0].lastState.terminated.reason

Alert attesi

kube-prometheus-stack include già le regole che scattano su questi guasti, dopo la durata for: (5–15 minuti):

Alert	Componenti
KubePodCrashLooping	api
KubePodNotReady	frontend, worker, checkout, orders-db
KubeDeploymentReplicasMismatch	tutti i deployment rotti
KubeContainerWaiting	frontend, worker

curl -s localhost:9093/api/v2/alerts | jq -r '.[] | select(.labels.namespace=="shop-demo") | "\(.labels.alertname)\t\(.labels.pod // .labels.deployment)"'

curl -s 'localhost:9090/api/v1/query' --data-urlencode 'query=ALERTS{namespace="shop-demo",alertstate="firing"}' | jq -r '.data.result[].metric | "\(.alertname) \(.pod // .deployment // "")"'