This sample demonstrates the workaround for problem when a namespace is being deleted with a running controller, that watches resources in its own namespace. If the pod or other underlying resources (role, role binding, service account) are deleted before the cleanup of the custom resource the namespace deletion is stuck.
see also: #2528