RFE: Watch labels

[squeed]

Feature Request

In #314 / #428, the ability was added to watch by label selector. This is useful if the watch would otherwise cause a lot of unnecessary objects to be kept in cache / sent over the wire. However, with the move to the controller-runtime, it seems that feature was lost.

What I really need for my particular use case is the ability to watch for all configmaps with a specific label, in all namespaces. I'd like to avoid keeping all configmaps in memory.

It may be that this particular use-case is too far outside the scope of a normal operator, in which case that's fine. I still think that label-based watches would be useful :-)

[joelanford]

I'm not sure if @hasbro17 has any background on this, but this seems do-able as a PR to controller-runtime.

I did my best to trace back from the source.Kind to see how it populates the cache, and it appears that client-go's cache.Reflector is what actually makes the List and Watch() calls to the API using the cache.ListWatch setup by controller-runtime here and here. However controller-runtime currently doesn't expose a way to modify the ListOptions.

Perhaps it would be possible to update source.Kind to use NewFilteredListWatchFromClient() (or some variant of that) and expose optionsModifier as a public field on source.Kind?

[hasbro17]

@joelanford There's already a feature request for this upstream: kubernetes-sigs/controller-runtime#244

I think the general idea is that we'll want to pass down a predicate to a controller's Watch that would dynamically make the informer's ListWatch filtered.
However that leaves you with the issue of the default cached client now using a filtered cache.
There's some options for dealing with that but that's something we need to figure out upstream with a design proposal/poc.

[openshift-bot]

Issues go stale after 90d of inactivity.

Mark the issue as fresh by commenting /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.
Exclude this issue from closing by commenting /lifecycle frozen.

If this issue is safe to close now please do so with /close.

/lifecycle stale

[openshift-bot]

Stale issues rot after 30d of inactivity.

Mark the issue as fresh by commenting /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.
Exclude this issue from closing by commenting /lifecycle frozen.

If this issue is safe to close now please do so with /close.

/lifecycle rotten
/remove-lifecycle stale

[openshift-bot]

Rotten issues close after 30d of inactivity.

Reopen the issue by commenting /reopen.
Mark the issue as fresh by commenting /remove-lifecycle rotten.
Exclude this issue from closing again by commenting /lifecycle frozen.

/close

[openshift-ci-robot]

@openshift-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.

Reopen the issue by commenting /reopen.
Mark the issue as fresh by commenting /remove-lifecycle rotten.
Exclude this issue from closing again by commenting /lifecycle frozen.

/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.