KEP 1287: Graduate InPlacePodVerticalScaling to GA

[natasha41575]

• One-line PR description: Graduate InPlacePodVerticalScaling to GA.

• Issue link: In-Place Update of Pod Resources #1287

• Other comments:

We believe we are ready to graduate InPlacePodVerticalScaling to GA in kubernetes 1.35.

Per the graduation requirements:

• VPA integration has moved to beta.
• No major bugs have been reported for 3 months.
• Pod-scoped resources will be handled by this enhancement, and we expect this to land in v1.35.
• UpdatePodSandboxResources has been implemented in crio, and is slated to land in the containerd 2.2 release in November (PR link). This is not the critical piece of functionality for in-place pod resize, so while we do want and expect it to land, we do not consider it blocking.
• Re-evaluation of the following decisions:
  • Resize atomicity - Resizes will stay atomic. Allowing partial resizes adds significant complexity and the use case is unclear.
  • Exposing allocated resources in the pod status - We will continue to expose allocated resources in the pod status.
  • QOS class changes - This is still considered out of scope for this KEP.

Most importantly, user feedback is green:

• Internal usage metrics show that usage is high. There have been no major bugs reported.
• There are now multiple integrations with in-place pod resize including Cast.ai, ScaleOps, and VPA. Scale-ops has even announced that their in-place pod resize integration is GA: https://scaleops.com/blog/now-ga-scaleops-launches-in-place-pod-resizing/.
• We have feedback from a user of Cast.ai that they are very happy with it.
• We have a customer that ran benchmarking of their workloads, and showed that they have 30% cost efficiency when using in-place pod resize vs without it.
• ...and a few others.

/sig node autoscaling scheduling

/cc @tallclair @dchen1107

/cc @dom4ha
for sig-scheduling

/cc @jackfrancis
for sig-autoscaling

[natasha41575]

We have a tracking bug for the Conformance endpoints: kubernetes/kubernetes#133607

[dom4ha]

/label lead-opted-in
/milestone v1.35

[natasha41575]

/cc @wojtek-t

[jpbetz]

/approve
For PRR (Everything appears to have been filled out by Beta, which is what I like to see!)

[tallclair]

/lgtm

[tallclair]

Oops, thanks for cleaning this up.

[jackfrancis]

/lgtm

for SIG Autoscaling

cc @adrianmoisey

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: jackfrancis, jpbetz, natasha41575
Once this PR has been reviewed and has the lgtm label, please assign mrunalp for approval. For more information see the Code Review Process.

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

Needs approval from an approver in each of these files:

• keps/prod-readiness/OWNERS [jpbetz]
• keps/sig-node/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment