New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add a blog post about decoupled taint eviction controller #43676

Merged

k8s-ci-robot merged 1 commit into kubernetes:main from yuanchen8911:taintmanager

Dec 13, 2023

Member

yuanchen8911 commented Oct 24, 2023 •

edited by sftim

Loading

Add a blog for KEP 3902

k8s-ci-robot added cncf-cla: yes size/XS labels

k8s-ci-robot requested a review from nate-double-u

October 24, 2023 21:26

k8s-ci-robot added the area/blog label

k8s-ci-robot requested a review from onlydole

October 24, 2023 21:26

k8s-ci-robot added language/en sig/docs labels

netlify bot commented Oct 24, 2023 •

edited

Loading

✅ Pull request preview available for checking

Built without sensitive environment variables

Name	Link
🔨 Latest commit	`fe882f1`
🔍 Latest deploy log	https://app.netlify.com/sites/kubernetes-io-main-staging/deploys/65678d567d6b910008613adf
😎 Deploy Preview	https://deploy-preview-43676--kubernetes-io-main-staging.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

To edit notification comments on pull requests, go to your Netlify site configuration.

yuanchen8911 mentioned this pull request

Decouple TaintManager from NodeLifecycleController kubernetes/enhancements#3902

Open

10 tasks

yuanchen8911 changed the title ~~Add a log post to decoupled taint eviction manager (placeholder)~~ Add a blog post to decoupled taint eviction manager (placeholder)

Contributor

sftim commented Oct 24, 2023

/retitle [WIP] Add a blog post about decoupled taint eviction manager

k8s-ci-robot changed the title ~~Add a blog post to decoupled taint eviction manager (placeholder)~~ [WIP] Add a blog post about decoupled taint eviction manager

k8s-ci-robot added the do-not-merge/work-in-progress label

Contributor

krol3 commented Nov 15, 2023 •

edited

Loading

Hi @yuanchen8911 here Communication Team 1.29, the deadline to the feature blog be ready to review is this Friday, Nov 17th, the proposal publish date will be Dec 19th. cc: @a-mccarthy @kcmartin @James-Quigley

@kubernetes/sig-docs-blog-owners: Blog scheduled: Dec 19th, Publication Order Nro:8

yuanchen8911 force-pushed the taintmanager branch from 02e7f10 to 8d35bce Compare

November 15, 2023 22:18

k8s-ci-robot added size/M and removed size/XS labels

yuanchen8911 changed the title ~~[WIP] Add a blog post about decoupled taint eviction manager~~ [WIP] Add a blog post about decoupled taint eviction controller (WIP)

yuanchen8911 force-pushed the taintmanager branch from 8d35bce to 80b263f Compare

November 15, 2023 22:39

yuanchen8911 changed the title ~~[WIP] Add a blog post about decoupled taint eviction controller (WIP)~~ [WIP] Add a blog post about decoupled taint eviction controller

yuanchen8911 force-pushed the taintmanager branch from 80b263f to 7ad71cb Compare

November 15, 2023 22:42

Member Author

yuanchen8911 commented Nov 15, 2023 •

edited

Loading

tengqm reviewed

View reviewed changes

content/en/blog/_posts/2023-10-24-taint-eviction-controller.md Outdated


		It may slightly increase the communication overhead from applying node taints to performing pod eviction.

		Will enabling/using this feature result in a non-negligible increase in resource usage (CPU, RAM, disk, IO, ...) in any components?

Contributor

tengqm Nov 16, 2023

no-neglisible ... can we avoid this double negating term?

Member Author

yuanchen8911 Nov 16, 2023

Removed the word non-negligible.

yuanchen8911 force-pushed the taintmanager branch from 7ad71cb to 114621e Compare

November 16, 2023 04:09

yuanchen8911 requested a review from tengqm

November 16, 2023 04:11

atosatto reviewed

View reviewed changes

content/en/blog/_posts/2023-10-24-taint-eviction-controller.md Outdated

Comment on lines 27 to 30

+              A new feature gate, `SeparateTaintManager`, has been added. To enable the new feature, users can use the following flags:
+              - kube-apiserver: `--feature-gates=SeparateTaintManager=true`
+              - kube-controller-manager: `--controllers=kube-eviction-controller`

Contributor

atosatto Nov 16, 2023

This is actually going to be enabled by default as the SeparateTaintManager feature-flag is in beta.

Member Author

yuanchen8911 Nov 18, 2023

Clarified it.

atosatto reviewed

View reviewed changes

content/en/blog/_posts/2023-10-24-taint-eviction-controller.md Outdated

+              - kube-apiserver: `--feature-gates=SeparateTaintManager=true`
+              - kube-controller-manager: `--controllers=kube-eviction-controller`
+              For compatibility, the legacy `TaintManager` can still be used with the `--taint-manager` flag.

Contributor

atosatto Nov 16, 2023

This is not correct. If users want to use the old taint-manager they need to set the feature-flag to false (e.g. SeparateTaintManager=false).

What this feature gives the users, is the ability to disable taint-based eviction by setting SeparateTaintManager=true and --controllers=-kube-eviction-controller.

Member Author

yuanchen8911 Nov 18, 2023

Updated it.

atosatto reviewed

View reviewed changes

content/en/blog/_posts/2023-10-24-taint-eviction-controller.md Outdated

+              - Adding a pre-defined set of `NoExecute` taints to nodes based on the node conditions.
+              - Performing pod eviction on NoExecute taints
+              With the latest release, the taint-based eviction implementation has been moved out of the `node-lifecycle-controller` into a separate and independent component called `TaintEvictionController`. This separation aims to disentangle code, enhance code maintainability, and facilitate future extensions to either component.

Contributor

atosatto Nov 16, 2023 •

edited

Loading

I'd add something high-lighting the benefits and the potential use-cases where this could be useful, as done in the KEP. I'd also mention the new metrics which have been introduced with this change.

Member Author

yuanchen8911 Nov 18, 2023

Added the following

Use Cases

This new feature will allow cluster administrators to extend and enhance the default TaintEvictionController and even replace the default TaintEvictionController with a custom implementation to meet different needs, e.g., better supportof stateful workloads that use PersistentVolume on local disks.

Member Author

yuanchen8911 Nov 18, 2023

Added the metrics description as follows.

As part of the change, addtitonal metrics are introduced to monitor taint-based pod evictionis.

pod_deletion_duration_seconds measures th latency between the time when a taint effect has been activated for the Pod and its deletion via TaintEvictionController.
pod_deletions_total reports the total number of Pods deleted by TaintEvictionController since its start

atosatto reviewed

View reviewed changes

content/en/blog/_posts/2023-10-24-taint-eviction-controller.md Outdated

Comment on lines 40 to 42

		Will enabling/using this feature result in an increase in the time taken by any operations covered by existing SLIs/SLOs?

		It may slightly increase the communication overhead from applying node taints to performing pod eviction.

Contributor

atosatto Nov 16, 2023

I don't think this is correct, as setting taints and evicting Pods is already done via separated reconcilers in the old implementation. We strived to the changes to not impact Customers experience in any meaningful way.

Member Author

yuanchen8911 Nov 18, 2023

Changed it to No.

atosatto reviewed

View reviewed changes

content/en/blog/_posts/2023-10-24-taint-eviction-controller.md Outdated


		As with any Kubernetes feature, multiple community members have contributed, from writing the KEP to implementing the new controller and reviewing the KEP and code. Special thanks to:

		- Aldo Culquicondor (@alculquicondor)

Contributor

atosatto Nov 16, 2023

I'd like us to add here @atiratree and @logicalhan. Their feedback on the implementation, alongside @soltysh and @Huang-Wei is what made this change possible! Thank you 🙇‍♂️!

Member Author

yuanchen8911 Nov 18, 2023

Added the names.

yuanchen8911 force-pushed the taintmanager branch from c078374 to d7fbb67 Compare

November 27, 2023 17:25

yuanchen8911 requested a review from sftim

November 27, 2023 17:25

Contributor

Ritikaa96 commented Nov 28, 2023

/lgtm
Thanks @yuanchen8911

k8s-ci-robot assigned Ritikaa96

k8s-ci-robot added the lgtm label

atiratree reviewed

View reviewed changes

content/en/blog/_posts/2023-12-19-taint-eviction-controller.md Outdated

+              ## How to use the new feature?
+              A new feature gate, `SeparateTaintEvictionController`, has been added. The feature is enabled by default as Beta in Kubernetes 1.29.
+              Please refer to the [feature gate document](/content/en/docs/reference/command-line-tools-reference/feature-gates.md).

Member

atiratree Nov 28, 2023

I think this should reference the relative path on the website and not on github as suggested here #43676 (comment)

Contributor

sftim Nov 28, 2023

Good point.

/lgtm cancel

Member Author

yuanchen8911 Nov 29, 2023

Replaced it with the short URL.

k8s-ci-robot removed the lgtm label

yuanchen8911 force-pushed the taintmanager branch from d7fbb67 to f021c1b Compare

November 28, 2023 14:47

yuanchen8911 requested a review from atiratree

November 28, 2023 14:47

Contributor

sftim commented Nov 28, 2023

Thanks

/lgtm

k8s-ci-robot added the lgtm label

Ritikaa96 suggested changes

View reviewed changes

content/en/blog/_posts/2023-12-19-taint-eviction-controller.md Outdated Show resolved Hide resolved

k8s-ci-robot removed the lgtm label

k8s-ci-robot requested a review from Ritikaa96

November 29, 2023 19:12


          Update taint-eviction-controller blog

fe882f1

Add authors

Update

Fix errors and address Andrea's comments

Removed duplication

Update 2023-10-24-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Update 2023-10-24-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Update 2023-10-24-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Update 2023-10-24-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Update 2023-10-24-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Address comment and rename the file

Fix format

Address comments

Add the feature gate reference

Fix typos and format issues

Fix reference link error

Fix a borken link

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

node to Node

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Tim Bannister <[email protected]>

Fix the reference link

Update content/en/blog/_posts/2023-12-19-taint-eviction-controller.md

Co-authored-by: Ritika <[email protected]>

yuanchen8911 force-pushed the taintmanager branch from baa5038 to fe882f1 Compare

November 29, 2023 19:13

Contributor

sftim commented Nov 30, 2023

/lgtm

k8s-ci-robot added the lgtm label

Contributor

sftim commented Dec 1, 2023

/lgtm cancel
/approve cancel

Requires new date (release of v1.29 has been postponed)

k8s-ci-robot removed lgtm approved labels

Contributor

sftim commented Dec 3, 2023

The new publication date is 2023-12-19; please adjust the article.

Contributor

sftim commented Dec 4, 2023

Thanks
/lgtm
/approve

Do not merge until Kubernetes v1.29 is released.

k8s-ci-robot added the lgtm label

Contributor

k8s-ci-robot commented Dec 4, 2023

LGTM label has been added.

Git tree hash: 2b872fda0ba6c964d2b70dc06dc0f42758e71c56

Contributor

k8s-ci-robot commented Dec 4, 2023

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: sftim

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:

~~content/en/blog/OWNERS~~ [sftim]

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

k8s-ci-robot added the approved label

Member

reylejano commented Dec 13, 2023

1.29 is released
/unhold

k8s-ci-robot removed the do-not-merge/hold label

k8s-ci-robot merged commit 751ec13 into kubernetes:main

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

nate-double-u Awaiting requested review from nate-double-u

onlydole Awaiting requested review from onlydole

atosatto Awaiting requested review from atosatto

tengqm Awaiting requested review from tengqm

a-mccarthy Awaiting requested review from a-mccarthy

sftim Awaiting requested review from sftim

atiratree Awaiting requested review from atiratree

Ritikaa96 Awaiting requested review from Ritikaa96

Labels

approved area/blog cncf-cla: yes language/en lgtm sig/docs size/M