Cherry-pick #15067 endpoints fixes.

[joelsmith]

cc @sjenning @eparis @derekwaynecarr

This is a cherry-pick of #15067 to prevent too-frequent processing of endpoint updates.

https://bugzilla.redhat.com/show_bug.cgi?id=1481603

Here's the original text of #15067 as written by @aveshagarwal:

Fixes #14710

Note: I had to manually merge changes and due to this, the commits might not match exactly.

It has following 2 complete upstream PRs:
kubernetes/kubernetes#47731 (with some modifications/additions).
kubernetes/kubernetes#47788

And one partial PR to make above PRs compile:
kubernetes/kubernetes#44408

(Edited to fix bz link -- originally sent with wrong bug)

[deads2k]

/unassign
/assign eparis

[aveshagarwal]

There has been a PR for this already #15343. I think I will close it now.

[smarterclayton]

Still seeing excessive write rates on endpoints on 3.6.1 #16583

[joelsmith]

@smarterclayton this patch should decrease updates for endpoints, although only as a byproduct.

In my observations, the periodic re-checking of every service resulted in the non-update updates getting pushed out whenever a pod changed. With this fix, the endpoint controller will decide whether to update the endpoint whenever a pod event occurs. Without it, every 30 seconds every endpoint gets enqueued for sync, and that results in the non-update updates (i.e., changes only in the resource version of the target ref pod) being pushed out.

[derekwaynecarr]

/kind bug
/approve
/lgtm

[openshift-merge-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: derekwaynecarr

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these OWNERS Files:

• OWNERS [derekwaynecarr]

You can indicate your approval by writing /approve in a comment
You can cancel your approval by writing /approve cancel in a comment

[openshift-merge-robot]

Automatic merge from submit-queue.