…roxy#1860)

This reverts commit f47be07.

Signed-off-by: JP Simard <jp@jpsim.com>

Signed-off-by: JP Simard <jp@jpsim.com>

…out inprogress (#1860)

**Description**
There have been sequences when the routes have been deleted and created.
The extProc sidecar would not be injected. We noticed this due to a
condition when the rollout is in progress, and a few pods are in a
terminating state but have an extProc sidecar in them. It does not
inject the sidecar in some cases (e.g., with multiple replicas).

To fix this issue, we add a check to look for deployment/daemonset
status if they are in progress. Requeue with a 5-second timeout.
Additionally, check whether the sidecar is present in each pod. This is
to ensure that, in multi-replica cases, there are no scenarios with
inconsistent state (some with sidecar and some without). If there is an
inconsistent state, we force rollout.

Changes:
- Skip terminating pods from checks and Consolidated checking of pods
with sidecar into `checkPodHasSideCar()`
- Included `isRolloutInProgress()` for checking inprogress


**Related Issues/PRs (if applicable)**
Fixes: #1857 

**Special notes for reviewers (if applicable)**

Tested this change by running a sequence of creation -> deletion of
routes. After the sequence was complete. The replicas were inconsistent
state which triggered a force rollout

```
INFO	controller.gateway	pods are inconsistent while rollout is stable, forcing rollout	{"podsWithSidecarSeen": true, "podsWithoutSidecarSeen": true}
```

Requesting thoughts/feedback on this approach

---------

Signed-off-by: Gavrish Prabhu <gavrish.prabhu@nutanix.com>