Failing Pods

Failing Pods

Diagnosing a Failing BookNest Pod

Work outside in: get, then describe and events, then logs, then debug. An API canary asking for two CPUs fails before any Pod exists:

A Deployment with no Pods at allShell

kubectl create deployment api-canary --image=localhost:33500/booknest-api:1.4 \
  --dry-run=client -o yaml \
  | yq '.spec.template.spec.containers[0].resources.requests.cpu = "2"' \
  | kubectl apply -f - >/dev/null 2>&1
sleep 5; kubectl get deployment api-canary; kubectl get pods -l app=api-canary
kubectl events --for=replicaset/$(kubectl get rs -l app=api-canary -o name | cut -d/ -f2) \
  | grep -m1 -o 'exceeded quota: .*' | fold -s -w 95
kubectl delete deployment api-canary >/dev/null

Output

NAME         READY   UP-TO-DATE   AVAILABLE   AGE
api-canary   0/1     0            0           5s
No resources found in booknest namespace.
exceeded quota: booknest, requested: requests.cpu=2, used: requests.cpu=520m, limited:
requests.cpu=1

The quota of Namespaces and Labels rejected every Pod at admission; only the ReplicaSet's FailedCreate event says so. A Pending Pod points to the scheduler (FailedScheduling), one Running but not Ready to its readiness probe (Unhealthy).

MENU