A control-plane upgrade finishes and later policy behavior differs by admission path for the same resource kind.
A Gatekeeper policy looks unchanged and one namespace starts failing admission
The policy object is present, yet one control-plane path evaluates an earlier schema version and rejects resources the new template should allow.
Scenario
What to check first
- Identify the primary failure signal in the The Policy Was the Same and One API Server Brought Yesterday's Schema to the Meeting scenario.
- Separate visible symptoms from the underlying technical dependency.
- Describe the safest recovery path and the follow-up prevention work.
Checking checklist
- Summarize the current impact and the last known change.
- Collect direct evidence from logs, runtime state, and configuration before changing anything.
- Separate immediate recovery from permanent prevention work.
Recovery and prevention
Inspect CRD conversion behavior per API server before editing constraints or disabling Gatekeeper.
Questions worth viewing together
Community-field Kubernetes problem inspired by admission-control discussions where one API server returned a stale Gatekeeper schema. Admission drift can hide in CRD conversion and schema-serving paths across partially upgraded control planes.
Teams often blame policy logic when one API server is still converting against an older template schema.
Rolling control-plane upgrades should validate CRD conversion webhooks and schema parity across every API server.
Similar cases seen in the field