Readiness failure removes Pod from Service Endpoints
Aggressive liveness probes cause CrashLoopBackOff
Wrong readiness path can make a healthy app receive no traffic
ConfigMaps and Secrets
Symptom
Cause
CreateContainerConfigError
Missing ConfigMap/Secret/key
File missing in mounted path
Wrong mountPath, missing subPath, or key name mismatch
Env var absent
envFrom / secretKeyRef points to wrong object/key
NetworkPolicy
Default is allow until a policy selects Pods
podSelector: {} selects all Pods
Empty ingress: [] denies all ingress for selected Pods
Egress default-deny breaks DNS unless UDP/TCP 53 is allowed
Multiple from entries are OR; multiple selectors in one entry are AND
Storage
Symptom
Check
PVC Pending
kubectl describe pvc, StorageClass name
Mount failure
PVC bound? claimName correct?
StatefulSet data remains
PVCs are not deleted with StatefulSet
Commands that save time
kubectl describe pod <pod>
kubectl logs <pod>--previous
kubectl get events --sort-by=.metadata.creationTimestamp
kubectl label pod <pod>key=value --overwrite
kubectl get <resource>-o yaml
Final checklist
Practiced SA → Role → RoleBinding 5x
Can fix Service 0 Endpoints in under 2 min
Know port vs targetPort vs Ingress backend port
Can create a CronJob and trigger it with --from=cronjob
Can explain NetworkPolicy selector logic
Can debug ImagePullBackOff, CrashLoopBackOff, and CreateContainerConfigError