fikaworks / kubectl-plugins Goto Github PK

This will allow people with multiple clusters to more easily use these plugins

As a developer
I want to be able to add label selectors to the prune to limit the scope of the command
In order to have more control over which secrets and configmaps get pruned.

Background

We use Kustomize and the configmapgenerator. It leaves behind quite a few unused configmaps, so using the plugin should help us a lot to mitigate this. However, the tool currently doesn't allow us to limit the prune to the configmaps that we are actively managing. We use some operators and one of those creates a configmap which is used by the operator directly through API calls and is not used in any other configuration. With labels, we could limit the prune to just the configmaps that are generated with Kustomize.

Most people version ConfigMaps/Secrets by creating new ones instead of editing existing. This pattern is heavily employed by tools like kustomize, as it is a best practice.

When you roll back a deployment (i.e. activating a replicaset that was scaled down to 0 pods) you don't want to find out you've pruned configmap/secret it referenced...

I think this plugin should look at all ReplicaSet specs for such unused info. If there's a ReplicaSet spec not cleaned up by its Deployment object (due to max history), its secret/cm should not be deleted.

I get nervous when there's no dry-run option on this.

Hey, thanks for the great tool. I'm wondering how difficult it would be to support passing multiple cronjobs to exec, so we could use the plugin like:

kubectl exec-cronjob some-cj-{a,b,c}

Thanks again!