KEP-4951: Configurable tolerance for HPA

[jm-franc]

Add KEP-4951: Configurable tolerance for HPA

Issue link: #4951

[k8s-ci-robot]

Welcome @jm-franc!

It looks like this is your first PR to kubernetes/enhancements 🎉. Please refer to our pull request process documentation to help your PR have a smooth ride to approval.

You will be prompted by a bot to use commands during the review process. Do not be afraid to follow the prompts! It is okay to experiment. Here is the bot commands documentation.

You can also check if kubernetes/enhancements has its own contribution guidelines.

You may want to refer to our testing guide if you run into trouble with your tests not passing.

If you are having difficulty getting your pull request seen, please follow the recommended escalation practices. Also, for tips and tricks in the contribution process you may want to read the Kubernetes contributor cheat sheet. We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to Kubernetes. 😃

[k8s-ci-robot]

Hi @jm-franc. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[raywainman]

This is great! Looking forward to tackling a longstanding user request! :)

[pr00se]

Excellent work @jm-franc, thank you!

I left some suggestions in the comments, feel free to use them or tweak them further. I skipped the sections that are needed for Beta, since we're still a ways off (even the Alpha ones may be premature).

I think we can merge the end result as-is, and then work with the sig to get it to the implementable state.

[jm-franc]

Excellent work @jm-franc, thank you!

I left some suggestions in the comments, feel free to use them or tweak them further. I skipped the sections that are needed for Beta, since we're still a ways off (even the Alpha ones may be premature).

I think we can merge the end result as-is, and then work with the sig to get it to the implementable state.

Thank you for all those suggestions Patryk, I've merged all of them. I'm now sending this KEP for review.

[jm-franc]

/assign @gjtempleton

Assigning to Guy for review/approval.

[raywainman]

/lgtm

[raywainman]

Small ask - I wonder if we can emit a message in the HPA status when the tolerance level is preventing a scale up or scale down?

(looking at the code, it might be tricky to add but wonder what your thoughts are and whether it would be useful)

[jm-franc]

I looked it up, but I can't find any straightforward way to do this.

1. We don't want to emit a message when the HPAs is just in-tolerance (that's almost always the case), but we could emit it only when the computed number of replicas is different from the current one.
2. The message would still trigger frequently for metrics valued just at the boundary between 2 replicas.
3. It's not clear what to do when some metrics are in tolerance while others aren't (and even if some are in-tolerance, it doesn't mean that a smaller tolerance would change the final recommendation).

This doesn't look unsolvable: we could for example compute the recommendation both normally and with tolerances set to 0, then warn users if the results are different. (But this would be a large refactoring).

If I'm missing a simpler solution I would be all in favour for this. If it's as complicated as I currently see this, I'd keep this outside of this KEP as this is only tangentially related to the problem addressed here.

[raywainman]

Completely reasonable. Thanks for checking.

[k8s-ci-robot]

New changes are detected. LGTM label has been removed.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: jm-franc
Once this PR has been reviewed and has the lgtm label, please ask for approval from gjtempleton. For more information see the Kubernetes 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/sig-autoscaling/OWNERS

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

[jm-franc]

/retest

[k8s-ci-robot]

@jm-franc: Cannot trigger testing until a trusted user reviews the PR and leaves an /ok-to-test message.

In response to this:

/retest

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[raywainman]

/ok-to-test

[jm-franc]

/retest