When the Pod Is Fine but Nothing Works — Debugging Kubernetes Services and PVCs
Last time I deliberately broke Pods three ways and watched them fail to start
(CrashLoopBackOff and friends).
This time the Pods are all Running and healthy — and things still don't work.
The bug isn't in the Pod; it's in the plumbing around it: the Service → Endpoints → Pod
forwarding chain, and the PVC → StorageClass → PV binding chain.
The skill here is knowing which link broke without guessing.
