Docker & KubernetesCrashLoopBackOff thường là app crash hoặc liveness probe giết container; xem logs hiện tại/previous và events. ImagePullBackOff thường do image tag sai, registry auth, network hoặc rate limit. Pending thường do thiếu resources, PVC chưa bind, node selector/taints hoặc quota.
Checklist: kubectl describe pod, kubectl logs --previous, kiểm image/tag/secret, requests/limits, events namespace, PVC, node capacity và rollout history. Không chỉ restart Pod nếu chưa hiểu root cause.
CrashLoopBackOff usually means the app crashes or a liveness probe kills the container; inspect current/previous logs and events. ImagePullBackOff often comes from a wrong image tag, registry auth, network or rate limit. Pending often comes from insufficient resources, unbound PVC, node selector/taints or quota.
Checklist: kubectl describe pod, kubectl logs --previous, image/tag/secret, requests/limits, namespace events, PVC, node capacity and rollout history. Do not just restart Pods without understanding root cause.