TIL: deleting a CRD deletes every instance of that kind across the cluster

I used to treat CRD deletion like uninstalling an operator binary — remove the type definition, leave the objects alone, clean up later. Kubernetes doesn’t work that way.

When you delete a CustomResourceDefinition, the API server garbage-collects all custom resources of that kind in every namespace. Not just the ones in the same namespace as the CRD (CRDs are cluster-scoped anyway). The whole inventory for that API group/kind goes with it.

kubectl delete crd certificates.cert-manager.io
# every Certificate object in every namespace is gone too

That’s by design: without the CRD, the API server has no schema to store or serve those objects under. So “delete the CRD to reinstall cleanly” is a cluster-wide data wipe for that kind, not a local reset.

Same trap shows up when Helm or an operator chart owns the CRD and an uninstall/upgrade removes it — the instances leave with the definition. Before you kubectl delete crd …, assume every live object of that kind is about to disappear.