KEP: Surge capacity on eviction instead of blocking on PodDisruptionBudget

[monta-riahi]

What this PR does

Adds a provisional KEP proposing an opt-in behavior where an exhausted PodDisruptionBudget triggers a surge of replacement capacity before an eviction completes, instead of blocking the eviction outright with 429 TooManyRequests.

Today API-initiated eviction is a binary decision against the PDB: if DisruptionsAllowed == 0 the eviction is rejected indefinitely. For a single-replica Deployment with minAvailable: 1 this makes the workload undrainable without an outage. This KEP proposes bringing up a replacement first (like maxSurge in a rolling update) and releasing the eviction once the replacement is Ready.

The document is intentionally provisional: it frames the problem, lays out the three cooperating pieces (signal → surge → release), and enumerates the open design questions for SIG discussion (where surge intent lives, who owns the surge, non-surgeable workloads, and the async eviction contract). It is expected to compose with KEP-4212 Declarative Node Maintenance.

/sig node
/sig apps
/kind kep

[kubernetes-prow]

Welcome @monta-riahi!

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. 😃

[kubernetes-prow]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: monta-riahi
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.

Details

Needs approval from an approver in each of these files:

• 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

[kubernetes-prow]

Hi @monta-riahi. 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.

Details

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.

[monta-riahi]

Description from kubernetes/kubernetes #140599,

When K8s currently evicts a pod, the pod gets terminated/deleted and only then the controller will schedule a new replacement pod, this means that deployments with one replica will experience outage, unlike when rolling them out which will surge a new pod first and only then terminate the old one.

This is rather poor behavior, and it would be much nicer if K8s can be configured to block eviction until a new pod is scheduled (Surging the configured replicas count if needed) instead. e.g. Instead of a PodDisruptionBudget just outright blocking the eviction, it triggers a capacity surge to support the requested budget, allowing the eviction to continue once the new pods are available.