Merge pull request #16277 from kargakis/cherry-pick

Automatic merge from submit-queue

UPSTREAM: 52127: Fix deployment timeout reporting

If the previous condition has been a successful rollout then we
shouldn't try to estimate any progress. Scenario:

* progressDeadlineSeconds is smaller than the difference between
  now and the time the last rollout finished in the past.
* the creation of a new ReplicaSet triggers a resync of the
  Deployment prior to the cached copy of the Deployment getting
  updated with the status.condition that indicates the creation
   of the new ReplicaSet.

The Deployment will be resynced and eventually its Progressing
condition will catch up with the state of the world.

Signed-off-by: Michail Kargakis <mkargaki@redhat.com>

Cherry-pick of kubernetes/kubernetes#52127
/assign @mfojtik @tnozicka 
/area apps
/kind bug

Author: openshift-merge-robot