Ideally, stale resources are counted and automatically linked to a CI job to facilitate the ci-watcher's investigation and possibly open bugs automatically.

Also, resources that are deleted by prune because were leaked by cluster destroy command should be better analyzed.