KEP-961: Bump Statefulset MaxUnavailable to beta (built on top of #4474) #5228

Edwinhr716 · 2025-04-03T17:52:48Z

Supersedes #4474

One-line PR description: Bump Statefulset MaxUnavailable to beta

Issue link: maxUnavailable for StatefulSets #961

Other comments: kept @kerthcet's and @knelasevero's commits here. My commits: adding metric information

Metrics PR: kubernetes/kubernetes#130951

Signed-off-by: kerthcet <kerthcet@gmail.com>

Signed-off-by: Lucas Severo Alves <lseveroa@redhat.com>

Edwinhr716 · 2025-04-03T17:52:57Z

/assign @janetkuo

keps/sig-apps/961-maxunavailable-for-statefulset/README.md

janetkuo · 2025-04-03T21:25:10Z

keps/sig-apps/961-maxunavailable-for-statefulset/README.md

+  cluster required to make on upgrade, in order to maintain previous behavior?
+  - This is a new field, it will only be enabled after an upgrade. To maintain the
+    previous behavior, you can disable the feature gate manually in Beta or leave
+    the field unconfigured.


Perhaps we can discuss the case of when existing clusters already use this feature in alpha before upgrade.

Also, discuss what happens if the feature gate was disabled, and the field was set, before upgrading

Added both cases, PTAL

keps/sig-apps/961-maxunavailable-for-statefulset/README.md

janetkuo · 2025-04-03T22:15:51Z

keps/sig-apps/961-maxunavailable-for-statefulset/README.md

 If their StatefulSet rollingUpdate section has the field maxUnavailable specified with
-a value different than 1.
-The below command should show maxUnavailable value:
+a value different than 1. While in alpha and beta, the feature-gate needs to be enabled.


You can still set it to 1 when this feature is enabled, right? Perhaps they just need to check the existence of the maxUnavailable field, and that the feature gate is enabled.

true, it can be both. During rollout though you won't see a behavioral change if it is set to 1, but you will see it if it set to anything else. As an example, LWS sets the value even if the feature flag was disabled. The only way we could tell it wasn't enabled by default was by testing it with a value larger than one.

keps/sig-apps/961-maxunavailable-for-statefulset/README.md

keps/sig-apps/961-maxunavailable-for-statefulset/kep.yaml

Co-authored-by: Janet Kuo <chiachenk@google.com>

soltysh · 2025-06-05T10:55:08Z

/label tide/merge-method-squash

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

kannon92 · 2025-06-13T18:26:31Z

PRR shadow:

Taking a look before @wojtek-t.

@soltysh calls out here that there are a lot of TODO, bugs and metrics to add for alpha -> beta generation.

I even see unresolved threads about existing bugs that have blocked the promotion before. I see that PRs are up for these.

As part of https://github.com/kubernetes/enhancements/tree/master/keps/sig-architecture/5241-beta-featuregate-promotion-requirements#summary, a requirement for features to be promoted to beta is that metrics, features and bugs are all complete.

I think the PRR seems fine to me but I want to call out that all these items must be complete before the feature gate gets turned on.

My main concern is beta requirements. This seems like a pretty hefty list to achieve in a few weeks.

wojtek-t

Just two minor comments from me - other than that it LGTM from PRR pov and is waiting for SIG approval from @soltysh

keps/sig-apps/961-maxunavailable-for-statefulset/README.md

wojtek-t · 2025-06-16T08:54:45Z

keps/sig-apps/961-maxunavailable-for-statefulset/README.md

+
+Administrators should monitor the following metrics to assess the need for a rollback:
+
+- `statefulset_unavailability_violation`: Ths metric reflects the number of times the maxUnavailable condition is violated (i.e. spec.replicas - availableReplicas > maxUnavailable).


Please ensure that this metric will get added before enabling the feature by default.

soltysh · 2025-06-17T16:16:23Z

I think the PRR seems fine to me but I want to call out that all these items must be complete before the feature gate gets turned on.

I think that's the best path forward. I will personally ensure we DON'T enable the feature gate before all the missing bits pointed out here are addressed.

soltysh

/lgtm
/approve
from sig-apps pov

/hold
for @wojtek-t PRR approval

wojtek-t · 2025-06-17T17:58:29Z

The PRR looks fine - please just ensure that the following happen before FG is turned on to beta:

the metric is added as describe
the enablement test mentioned in the PRR
the manual test is run
[in addition to some other tests from tests section]

/approve PRR
/hold cancel

k8s-ci-robot · 2025-06-17T17:58:38Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: Edwinhr716, soltysh, wojtek-t

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

The pull request process is described here

Needs approval from an approver in each of these files:

~~keps/prod-readiness/OWNERS~~ [soltysh,wojtek-t]
~~keps/sig-apps/OWNERS~~ [soltysh]

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

wojtek-t · 2025-06-17T17:58:51Z

/lgtm

kerthcet and others added 10 commits April 3, 2025 17:13

Adapt to the latest template

4a1e825

Signed-off-by: kerthcet <kerthcet@gmail.com>

Bump to beta

4743543

Signed-off-by: kerthcet <kerthcet@gmail.com>

address comments

504b351

Signed-off-by: kerthcet <kerthcet@gmail.com>

Address commetns

a7aaaa9

Signed-off-by: kerthcet <kerthcet@gmail.com>

Add new metric rollingUpdateDurationSeconds to statefulset

22e0287

Signed-off-by: kerthcet <kerthcet@gmail.com>

go over review from kubernetes#3997, change metric

35c55fa

review notes, still need approval on metrics

aa89ff8

running and available instead of running and ready

3ccb4ef

review notes and UNRESOLVED blocks

6690da2

Signed-off-by: Lucas Severo Alves <lseveroa@redhat.com>

added metric information

36e0a64

k8s-ci-robot added the cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. label Apr 3, 2025

k8s-ci-robot requested review from kow3ns and palnabarun April 3, 2025 17:52

k8s-ci-robot added kind/kep Categorizes KEP tracking issues and PRs modifying the KEP directory sig/apps Categorizes an issue or PR as relevant to SIG Apps. labels Apr 3, 2025

github-project-automation bot added this to SIG Apps Apr 3, 2025

github-project-automation bot moved this to Needs Triage in SIG Apps Apr 3, 2025

k8s-ci-robot assigned janetkuo Apr 3, 2025

k8s-ci-robot added the size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. label Apr 3, 2025

janetkuo reviewed Apr 3, 2025

View reviewed changes

Edwinhr716 and others added 6 commits April 7, 2025 17:38

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

fc9798e

Co-authored-by: Janet Kuo <chiachenk@google.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

20d8a21

Co-authored-by: Janet Kuo <chiachenk@google.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

e818d14

Co-authored-by: Janet Kuo <chiachenk@google.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

af31565

Co-authored-by: Janet Kuo <chiachenk@google.com>

addressed comments

b42dfb6

added metric for detection

65b339b

soltysh mentioned this pull request Apr 17, 2025

maxUnavailable for StatefulSets #961

Open

8 tasks

k8s-ci-robot added the tide/merge-method-squash Denotes a PR that should be squashed by tide when it merges. label Jun 5, 2025

Edwinhr716 and others added 11 commits June 9, 2025 18:00

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

2e4eb62

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

613a867

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

2579caa

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

28d4614

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

d0b680d

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

33b1400

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

3a09deb

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

Update keps/sig-apps/961-maxunavailable-for-statefulset/README.md

34808bc

Co-authored-by: Maciej Szulik <soltysh@gmail.com>

addressing comments

e97a6d7

updated test coverage numbers

e320bb1

reworded sentences

1d729bf

Edwinhr716 mentioned this pull request Jun 14, 2025

Added types of podManagementPolicy to StatefulSet documentation kubernetes/website#51256

Open

wojtek-t reviewed Jun 16, 2025

View reviewed changes

soltysh approved these changes Jun 17, 2025

View reviewed changes

k8s-ci-robot added the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Jun 17, 2025

k8s-ci-robot assigned soltysh Jun 17, 2025

k8s-ci-robot added lgtm "Looks good to me", indicates that a PR is ready to be merged. approved Indicates a PR has been approved by an approver from all required OWNERS files. labels Jun 17, 2025

added explanation on mid-rollout flag disable

b37c46f

k8s-ci-robot removed the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Jun 17, 2025

k8s-ci-robot removed the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Jun 17, 2025

k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Jun 17, 2025

k8s-ci-robot merged commit 4a9783f into kubernetes:master Jun 17, 2025
4 checks passed

k8s-ci-robot added this to the v1.34 milestone Jun 17, 2025

github-project-automation bot moved this from Needs Triage to Done in SIG Apps Jun 17, 2025


		Administrators should monitor the following metrics to assess the need for a rollback:

		- `statefulset_unavailability_violation`: Ths metric reflects the number of times the maxUnavailable condition is violated (i.e. spec.replicas - availableReplicas > maxUnavailable).

KEP-961: Bump Statefulset MaxUnavailable to beta (built on top of #4474) #5228

KEP-961: Bump Statefulset MaxUnavailable to beta (built on top of #4474) #5228

Uh oh!

Conversation

Edwinhr716 commented Apr 3, 2025

Uh oh!

Edwinhr716 commented Apr 3, 2025

Uh oh!

Uh oh!

janetkuo Apr 3, 2025

Choose a reason for hiding this comment

Uh oh!

Edwinhr716 Apr 8, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

janetkuo Apr 3, 2025

Choose a reason for hiding this comment

Uh oh!

Edwinhr716 Apr 8, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

soltysh commented Jun 5, 2025

Uh oh!

kannon92 commented Jun 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

wojtek-t left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

wojtek-t Jun 16, 2025

Choose a reason for hiding this comment

Uh oh!

Edwinhr716 Jun 17, 2025

Choose a reason for hiding this comment

Uh oh!

soltysh commented Jun 17, 2025

Uh oh!

soltysh left a comment

Choose a reason for hiding this comment

Uh oh!

wojtek-t commented Jun 17, 2025

Uh oh!

k8s-ci-robot commented Jun 17, 2025

Uh oh!

wojtek-t commented Jun 17, 2025

Uh oh!

Uh oh!

Uh oh!

kannon92 commented Jun 13, 2025 •

edited

Loading