Automatic merge from submit-queue (batch tested with PRs 55394, 55412). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Adds e2e tests for Pod Priority and Preemption in Cluster Autoscaler
This PR adds e2e tests for Pod Priority and Preemption in Clucter Autoscaler:
- shouldn't scale up when expendable pod is created
- should scale up when non expendable pod is created
- shouldn't scale up when expendable pod is preempted
- should scale down when expendable pod is running
- shouldn't scale down when non expendable pod is running
- shouldn't scale up when expendable pod is created
- should scale up when non expendable pod is created
- shouldn't scale up when expendable pod is preempted
- should scale down when expendable pod is running
- shouldn't scale down when non expendable pod is running
Automatic merge from submit-queue (batch tested with PRs 55331, 55272, 55228, 49763, 55242). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
use versiond group clients from client-go
**What this PR does / why we need it**:
Some **Deprecated** group clients are still used, replace them with versioned group clients.
**Which issue this PR fixes**: fixes#49760
**Special notes for your reviewer**:
/assign @caesarxuchao
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Adds e2e tests for Node Autoprovisioning:
This PR adds e2e tests for Node Autoprovisioning: …
- should create new node if there is no node for node selector
Automatic merge from submit-queue (batch tested with PRs 55061, 55157, 55231). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Adds e2e tests for Node Autoprovisioning:
Adds e2e tests for Node Autoprovisioning:
- shouldn't add new node group if not needed
- shouldn't scale up if cores limit too low, should scale up after limit is changed
Automatic merge from submit-queue (batch tested with PRs 53645, 54734, 54586, 55015, 54688). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Fix Incorrect Scale Subresources and HPA e2e ScaleTargetRefs
The HPA e2es failed to actually set `apiVersion` on the created HPAs, which previous was ignored. Since the polymorphic scale client was merged, this behavior is no longer tolerated (it was never correct to begin with, but it accidentally worked).
Additionally, the `apps` resources have their own version of scale. Until `apps/v1beta1` and `apps/v1beta2` go away, we need to support those versions in the scale client.
Together, these broke some of the HPA e2es.
Fixes#54574
```release-note
NONE
```
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. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Use gcloud for enabling/disabling autoscaling in e2e tests
This removes temporary solution added in #28011 as it's no longer necessary. Should reduce flakes caused by not waiting for master restart after disabling autoscaling.
Automatic merge from submit-queue (batch tested with PRs 50223, 53205). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Create e2e tests for Custom Metrics - Stackdriver Adapter and HPA based on custom metrics from Stackdriver
**What this PR does / why we need it**:
- Add e2e test for Custom Metrics - Stackdriver Adapter
- Add 2e2 test for HPA based on custom metrics from Stackdriver
- Enable HorizontalPodAutoscalerUseRESTClients option
**Release note**:
```release-note
Horizontal pod autoscaler uses REST clients through the kube-aggregator instead of the legacy client through the API server proxy.
```
Automatic merge from submit-queue (batch tested with PRs 53567, 53197, 52944, 49593). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Clean up in `cluster_size_autoscaling.go`
**What this PR does / why we need it**:
Fix `golint` errors.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Disable autoscaling before removing autoscaled node pool
This is to prevent flakes due to API calls failing in AfterEach during master restart, which is triggered by deleting an autoscaled node pool. Adding disable call before deleting node pool should prevent this as we'll wait for master restart in disableAutoscaler function.
While it may be faster to wait after deletion of autoscaled node pools, this is less complex and will be easier to remove in the future when changing autoscaling setttings no longer triggers master restart.
Automatic merge from submit-queue (batch tested with PRs 52905, 52766). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>..
Refactor parsing cluster autoscaler status, add logging error
Minor improvements to autoscaling test suite and e2e framework.
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>..
should use time.Since instead of time.Now().Sub
**What this PR does / why we need it**:
should use time.Since instead of time.Now().Sub
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
NONE
**Special notes for your reviewer**:
NONE
**Release note**:
```release-note
```
NONE
Automatic merge from submit-queue (batch tested with PRs 52843, 52710, 52821, 52844). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>..
improve retrying logic when checking CA status
This should reduce the flake rate in cluster size autoscaling test suite.
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>..
In autoscaling integration test, use allocatable instead of capacity for node memory
This makes the remaining cluster autoscaling test (integration test of HPA and CA working together to scale up the cluster) use node allocatable resources when computing how much memory we need to consume in order to trigger scale up/prevent scale down. Follow up to #52650 as that one is already merging.
cc @wasylkowski
Automatic merge from submit-queue
Fix failing autoscaling test in GKE
This should fix `[sig-autoscaling] Cluster size autoscaling [Slow] should increase cluster size if pending pods are small and there is another node pool that is not autoscaled [Feature:ClusterSizeAutoscalingScaleUp]` by getting a list of nodes from GKE nodepool in a different way (filtering nodes by labels.) Currently, gcloud command used for it is failing, as we only have GKE node pool name in the test and not the actual MIG name.
Krzysztof Jastrzebski