Node Performance Benchmark

[dchen1107]

Description

Node Performance Benchmark

Progress Tracker

• Before Alpha: Initially proposed at Increase maximum pods per node kubernetes#23349 (comment)
  • Write and maintain draft quality doc:
    • Initial doc: https://drive.google.com/file/d/0B5b9gVHyDCGPa2dISk1xa1lmOTQ/view
    • Demo slide: https://drive.google.com/a/google.com/file/d/0B5b9gVHyDCGPYklzS1BuOEFlUzQ/view
  • Design Approval: @dchen1107
    • Identify shepherd: @dchen1107
    • Identify secondary/backup contact point: @Random-Liu
  • Write (code + tests + docs) then get them merged.
    • Performance related node e2e tests (density test and resource usage tests):
      • Add density (batch pods creation latency and resource) and resource performance tests to `test-e2e-node' kubernetes#29764
      • Add density (batch pods creation latency and resource) and resource performance tests to `test-e2e-node' built for Linux only kubernetes#30026
      • Add Time Series Data and Labels in Node density test kubernetes#30333
      • Add tag [benchmark] to node-e2e-test where performance limits are not verified kubernetes#30568
      • Add logging time series to benchmark test kubernetes#30786
      • add throughput in perf data and disable --cgroups-per-qos kubernetes#31185
    • Test Infrastructure (benchmark jenkins ci job):
      • WIP: Support benchmark in run_e2e.go and jenkins kubernetes#30425
      • Add benchmark to jenkins kubernetes#31039
      • Add job config for kubelet benchmark ci test test-infra#426
      • skip benchmark in jenkins serial test kubernetes#31164
    • Kubelet tracing:
      • WIP: add operation tracing in kubelet for performance analysis kubernetes#31280 (pending)

FEATURE_STATUS: IN_DEVELOPMENT

Docs

• Write user docs and get them merged in: https://github.com/kubernetes/contrib/blob/master/node-perf-dash/README.md
• User docs go into http://github.com/kubernetes/kubernetes.github.io.
• When the feature has user docs, please add a comment mentioning @kubernetes/docs.
• When you get LGTM, you can check this checkbox, and the reviewer will apply the "docs-complete" label.

[dchen1107]

Next step:

• Finishing performance tracing
• Publish the link to access the node performance test
• Enable the release process to generate the node performance benchmark

[dchen1107]

cc/ @aronchick

[janetkuo]

Do we need docs for this task?

[jaredbhatti]

Ping on docs. Any docs needs for this feature?

[jaredbhatti]

@dchen1107 Another ping on docs. Any PRs you can point me to?

[devin-donnelly]

This is apparently getting publicized in two places:

• A blog post here: https://docs.google.com/document/d/16RiXzmYBBcz1gcdxYi_GA5S9jdc37UZJb624U08O41w/edit
• and via a readme: https://github.com/coufon/contrib/blob/b065e60a23dd1ca630e26ebfeba010ad0b1cf851/node-perf-dash/README.md

We should also add this to our list of internal tools.

EDITTED by @dchen1107:

Readme at: https://github.com/kubernetes/contrib/blob/master/node-perf-dash/README.md

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

Prevent issues from auto-closing with an /lifecycle frozen comment.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or @fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten
/remove-lifecycle stale

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close