A constraint upgrade is rolled out and only one custom kind shows inconsistent admission results across clients.
A Gatekeeper constraint rejects only one object kind
Most objects are validated with the new template, yet one kind still flows through a stale schema path on one control-plane node.
Scenario
What to check first
- Identify the primary failure signal in the The Constraint Updated and One API Path Brought an Older 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
Compare conversion paths and API server cache refreshes before disabling the constraint or force-restarting every control-plane node.
Questions worth viewing together
Community-field Kubernetes problem inspired by Gatekeeper and control-plane discussions where one aggregated API path still served a stale schema. Admission inconsistencies can be limited to one versioned path or kind even when the controller rollout looked healthy.
Teams often blame CRD corruption when one cached conversion path is evaluating an older schema.
Constraint and CRD upgrades should be tested across every served version path, not just the default kubectl path.
Similar cases seen in the field