API discovery docs do not provide ETag information for clients

[smarterclayton]

We implemented "re-discovery" if the resource name doesn't match one we know:

kubectl get posd  # <-- triggers full rediscovery

1. Unfortunately, this means that having a typo here will now always take >2-5s to resolve (if you're on a slow connection, it could be minutes).
2. If you typo twice in a row, you download all discovery docs twice (the odds of this actually helping are almost zero)
3. We don't have etag or "if not modified" checks on these resources... so the user has to download them every time and they are not small
4. We get rate limited by the client (throttled client side) when we download these - which means when you have lots of groups we get even slower.

We need to be smarter, this is very annoying even when latency is 0.

[smarterclayton]

@kubernetes/sig-cli-bugs

[liggitt]

cc @kubernetes/sig-api-machinery-feature-requests

[sttts]

5. We have one discovery request per api group. It could be one, with one eTag.

[lavalamp]

/assign @mbohlool

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

Prevent issues from auto-closing with an /lifecycle frozen comment.

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

Send feedback to sig-testing, kubernetes/test-infra and/or @fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

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

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten
/remove-lifecycle stale

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close