HPA e2e tests specify 20 as requestSizeMillicores on resource-consumer
for making CPU workload to test HPA feature, and the value is
hard-coded as 20. That means the tests expect 2% CPU workload on every
resource-consumer processes, but actual CPU usage(4 - 6%) is over than
the expected usage on some environment. Then HPA scales many pods than
the test expected, and the test is failed.
To make these tests stable, this changes the value to 100 (10% CPU
workload on each process).
Previously, the HPA controller ignored APIVersion when resolving the
scale subresource for a kind, meaning if it was set incorrectly in the
HPA's scaleTargetRef, it would not matter. This was the case for
several of the HPA e2e tests.
Since the polymorphic scale client merged, APIVersion now matters. This
updates the HPA e2es to care about APIVersion, by passing kind as a full
GroupVersionKind, and not just a string.
Automatic merge from submit-queue (batch tested with PRs 46835, 46856)
Made tests that create Horizontal Pod Autoscaler delete it after they are done.
**What this PR does / why we need it**:
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes#46847
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 46835, 46856)
Made WaitForReplicas and EnsureDesiredReplicas use PollImmediate and improved logging.
**What this PR does / why we need it**: Most importantly, this results in better logging: timeout is logged at the level of the caller, not the helper function, helping debugging.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
