Diagnose a Kubernetes pod that is not Ready, stuck Terminating, CrashLoopBackOff, Pending, ImagePullBackOff, or restarting. Follows a fixed check order and stops at the first confirmed cause instead of guessing. Use when a pod will not become Ready, a rollout is stuck, a pod takes ~30s to delete, a container is OOMKilled or exits 143, or metrics disagree with observed behaviour. Do not use for cluster-wide outages affecting many unrelated workloads (use k8s-incident-triage) or for pre-deploy manifest review (use prod-readiness-gate).
