ci: Never cancel benchmark runs on master by ntucker · Pull Request #4211 · reactive/data-client

ntucker · 2026-10-05T23:15:07Z

Requested by Nathaniel · project thread

Before: the benchmark workflows (benchmark.yml, benchmark-react.yml, benchmark-spread.yml) share one concurrency group for master pushes with GitHub's default single pending slot. Three quick master pushes touching their paths cancel the middle run, which marks that master commit red.

After: master push runs queue (queue: max, up to 100 pending), so every push's benchmark runs and publishes in order. PR runs still cancel superseded ones.

How: queue accepts an expression, so each workflow picks the mode per event, alongside the existing cancel-in-progress expression (GitHub rejects queue: max only together with cancel-in-progress: true):

concurrency:
  group: ${{ github.event_name == 'push' && 'gh-pages-bench-node-push' || format('benchmark-node-{0}', github.head_ref) }}
  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
  queue: ${{ github.event_name == 'push' && 'max' || 'single' }}

Verified on real runs with a temporary probe workflow using the same block (since dropped from the branch): nine rapid branch pushes left one run in progress and the rest pending in order, none cancelled; on the PR, each new push cancelled both the pending and the in-progress run, the same as a control workflow with no queue key. The benchmark workflows themselves also accepted the block on this PR's runs. Updates the CI rule in .cursor/rules/ci-config.mdc.

🤖 Generated with Claude Code

https://claude.ai/code/session_01FbcU32urUKzbRdg672xQUG

Note

Low Risk
CI concurrency-only change for benchmark workflows; no application or release logic touched.

Overview
Benchmark GitHub Actions workflows now queue master push runs instead of letting the shared concurrency group cancel middle commits (which turned those checks red on npm scoring).

benchmark.yml, benchmark-react.yml, and benchmark-spread.yml add event-aware concurrency: queue: max on push (up to 100 pending, run in order) and queue: single on pull requests, paired with the existing cancel-in-progress only for PRs. Inline comments document why master runs must not be cancelled.

CI agent rules in .cursor/rules/ci-config.mdc (and the generated .claude/rules/ci-config.md) replace the note that benchmark push groups used the default single slot with this per-event queue / cancel-in-progress pattern for benchmark*.yml.

^{Reviewed by Cursor Bugbot for commit b641ccf. Bugbot is set up for automated code reviews on this repo. Configure here.}

Benchmark workflows share one concurrency group for master pushes. With GitHub's default single pending slot, a third quick push cancels the pending one and marks master red. Push runs now use queue: max; PR runs keep cancel-in-progress. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FbcU32urUKzbRdg672xQUG

changeset-bot · 2026-10-05T23:15:12Z

⚠️ No Changeset found

Latest commit: b641ccf

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

vercel · 2026-10-05T23:15:16Z

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
docs-site	Ready	Preview	Oct 5, 2026 11:19pm UTC

ntucker · 2026-10-05T23:15:40Z

Staff engineer (Cursor agent): LGTM at 494d9c2, assuming the temporary probe workflow comes out before merge.

This is the follow-up I left on #4205, and it's the smallest fix: the three benchmark workflows keep their shared push group (so gh-pages writes stay serialized and in order) and just stop dropping the middle pending run. The per-event queue expression mirrors the existing cancel-in-progress expression, so PR behavior is unchanged, and the CI rule in both .cursor/rules/ci-config.mdc and .claude/rules/ci-config.md now describes the real setup instead of calling out the gap.

Before merge:

Delete .github/workflows/zz-queue-probe.yml (and ideally squash the six TEMP queue expression probe commits away).
Keep it Ready only once the probe is out, so Bugbot and the benchmark runs check the final diff.

chatgpt-codex-connector · 2026-10-05T23:18:56Z

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-05T23:21:07.002634Z	`b641ccf`	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

ntucker · 2026-10-05T23:19:14Z

Done: the probe workflows and all TEMP commits are gone. The branch is now the single commit b641ccf, which was marked ready after the cleanup, so Bugbot and the benchmarks check the final diff. This change needs no changeset because it only touches CI.

Generated by Claude Code

github-actions

Benchmark Spread

Details

Benchmark suite	Current: `b641ccf`	Previous: `b4b502d`	Ratio
`setOneEntity in 10k entity store`	`154` ops/sec (`±0.97%`)	`155` ops/sec (`±0.95%`)	`1.01`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Benchmark React

Details

Benchmark suite	Current: `b641ccf`	Previous: `10e15cc`	Ratio
`data-client: getlist-100`	`238.1` ops/s (`± 3.5%`)	`173.93` ops/s (`± 5.0%`)	`0.73`
`data-client: getlist-500`	`74.63` ops/s (`± 3.5%`)	`53.76` ops/s (`± 4.8%`)	`0.72`
`data-client: update-entity`	`741.76` ops/s (`± 8.8%`)	`434.78` ops/s (`± 10.5%`)	`0.59`
`data-client: update-user`	`666.67` ops/s (`± 10.8%`)	`425.72` ops/s (`± 10.2%`)	`0.64`
`data-client: getlist-500-sorted`	`74.63` ops/s (`± 9.6%`)	`55.26` ops/s (`± 8.5%`)	`0.74`
`data-client: update-entity-sorted`	`666.67` ops/s (`± 6.9%`)	`400` ops/s (`± 7.0%`)	`0.60`
`data-client: update-entity-multi-view`	`666.67` ops/s (`± 9.9%`)	`408.33` ops/s (`± 7.5%`)	`0.61`
`data-client: list-detail-switch-10`	`24.57` ops/s (`± 7.5%`)	`12.71` ops/s (`± 7.5%`)	`0.52`
`data-client: update-user-10000`	`131.58` ops/s (`± 13.3%`)	`84.03` ops/s (`± 15.3%`)	`0.64`
`data-client: invalidate-and-resolve`	`65.79` ops/s (`± 6.4%`)	`45.98` ops/s (`± 4.7%`)	`0.70`
`data-client: unshift-item`	`408.33` ops/s (`± 4.7%`)	`263.16` ops/s (`± 4.6%`)	`0.64`
`data-client: delete-item`	`555.56` ops/s (`± 5.6%`)	`333.33` ops/s (`± 6.0%`)	`0.60`
`data-client: move-item`	`322.58` ops/s (`± 11.2%`)	`204.17` ops/s (`± 9.8%`)	`0.63`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Benchmark

Details

Benchmark suite	Current: `b641ccf`	Previous: `10e15cc`	Ratio
`normalizeLong`	`430` ops/sec (`±4.32%`)	`424` ops/sec (`±4.38%`)	`0.99`
`normalizeLong Values`	`384` ops/sec (`±1.16%`)	`385` ops/sec (`±0.30%`)	`1.00`
`normalizeLong Scalar`	`343` ops/sec (`±3.74%`)	`369` ops/sec (`±3.63%`)	`1.08`
`normalizeLong Scalar update`	`892` ops/sec (`±0.22%`)	`918` ops/sec (`±0.21%`)	`1.03`
`denormalizeLong`	`224` ops/sec (`±5.73%`)	`245` ops/sec (`±5.79%`)	`1.09`
`denormalizeLong Values`	`221` ops/sec (`±4.91%`)	`229` ops/sec (`±4.35%`)	`1.04`
`denormalizeLong donotcache`	`1063` ops/sec (`±0.67%`)	`992` ops/sec (`±1.01%`)	`0.93`
`denormalizeLong Values donotcache`	`755` ops/sec (`±0.22%`)	`752` ops/sec (`±0.22%`)	`1.00`
`denormalizeLong Scalar donotcache`	`1162` ops/sec (`±0.33%`)	`1029` ops/sec (`±0.24%`)	`0.89`
`denormalizeShort donotcache 500x`	`1380` ops/sec (`±0.10%`)	`1402` ops/sec (`±0.10%`)	`1.02`
`denormalizeShort 500x`	`598` ops/sec (`±6.52%`)	`639` ops/sec (`±6.73%`)	`1.07`
`denormalizeShort 500x withCache`	`7206` ops/sec (`±0.09%`)	`6531` ops/sec (`±0.21%`)	`0.91`
`queryShort 500x withCache`	`3395` ops/sec (`±0.06%`)	`3178` ops/sec (`±2.76%`)	`0.94`
`buildQueryKey All`	`53525` ops/sec (`±0.82%`)	`59344` ops/sec (`±1.21%`)	`1.11`
`query All withCache`	`6392` ops/sec (`±2.87%`)	`6251` ops/sec (`±2.07%`)	`0.98`
`denormalizeLong with mixin Entity`	`219` ops/sec (`±6.28%`)	`217` ops/sec (`±7.98%`)	`0.99`
`denormalizeLong withCache`	`7689` ops/sec (`±0.25%`)	`8085` ops/sec (`±0.26%`)	`1.05`
`denormalizeLong withCache (Scalar churn)`	`7546` ops/sec (`±0.85%`)	`8027` ops/sec (`±1.01%`)	`1.06`
`denormalizeLong Values withCache`	`6409` ops/sec (`±1.48%`)	`4995` ops/sec (`±1.81%`)	`0.78`
`denormalizeLong Scalar withCache`	`7326` ops/sec (`±0.73%`)	`6193` ops/sec (`±0.36%`)	`0.85`
`denormalizeLong Scalar update withCache`	`5527` ops/sec (`±0.54%`)	`3649` ops/sec (`±0.23%`)	`0.66`
`denormalizeLong All withCache`	`6386` ops/sec (`±0.24%`)	`6359` ops/sec (`±0.29%`)	`1.00`
`denormalizeLong Query-sorted withCache`	`6580` ops/sec (`±2.40%`)	`6428` ops/sec (`±2.11%`)	`0.98`
`denormalizeLongAndShort withEntityCacheOnly`	`1626` ops/sec (`±1.38%`)	`1742` ops/sec (`±0.27%`)	`1.07`
`denormalize bidirectional 50`	`4274` ops/sec (`±10.08%`)	`4510` ops/sec (`±12.09%`)	`1.06`
`denormalize bidirectional 50 donotcache`	`44762` ops/sec (`±0.23%`)	`41392` ops/sec (`±0.61%`)	`0.92`
`getResponse`	`5143` ops/sec (`±4.38%`)	`4474` ops/sec (`±3.73%`)	`0.87`
`getResponse (null)`	`8750694` ops/sec (`±25.32%`)	`10187553` ops/sec (`±0.57%`)	`1.16`
`getResponse (clear cache)`	`205` ops/sec (`±7.62%`)	`205` ops/sec (`±8.56%`)	`1`
`getSmallResponse`	`3942` ops/sec (`±0.92%`)	`3555` ops/sec (`±0.56%`)	`0.90`
`getSmallInferredResponse`	`3118` ops/sec (`±0.24%`)	`2825` ops/sec (`±0.10%`)	`0.91`
`getResponse Collection`	`3823` ops/sec (`±4.34%`)	`4545` ops/sec (`±2.96%`)	`1.19`
`get Collection`	`2449` ops/sec (`±0.22%`)	`3019` ops/sec (`±0.19%`)	`1.23`
`get Query-sorted`	`4636` ops/sec (`±1.37%`)	`5066` ops/sec (`±1.38%`)	`1.09`
`setLong`	`443` ops/sec (`±0.33%`)	`427` ops/sec (`±0.31%`)	`0.96`
`setLongWithMerge`	`252` ops/sec (`±0.50%`)	`253` ops/sec (`±0.22%`)	`1.00`
`setLongWithSimpleMerge`	`269` ops/sec (`±0.24%`)	`264` ops/sec (`±0.88%`)	`0.98`
`setSmallResponse 500x`	`892` ops/sec (`±1.48%`)	`934` ops/sec (`±1.54%`)	`1.05`
`setMany 50x one-per-row`	`138` ops/sec (`±0.66%`)	`144` ops/sec (`±0.51%`)	`1.04`
`setMany 50 batch`	`3613` ops/sec (`±0.62%`)	`3609` ops/sec (`±0.48%`)	`1.00`
`setMany 500x one-per-row`	`14.2` ops/sec (`±0.83%`)	`14.49` ops/sec (`±0.59%`)	`1.02`
`setMany 500 batch`	`1464` ops/sec (`±3.53%`)	`1426` ops/sec (`±0.21%`)	`0.97`

This comment was automatically generated by workflow using github-action-benchmark.

ntucker self-assigned this Oct 5, 2026

ntucker force-pushed the claude/project-thread-4fyw11 branch from cf10c31 to b641ccf Compare October 5, 2026 23:18

ntucker marked this pull request as ready for review October 5, 2026 23:18

vercel Bot deployed to Preview October 5, 2026 23:19 View deployment

github-actions Bot reviewed Oct 5, 2026

View reviewed changes

ntucker merged commit 0ca688f into master Oct 5, 2026
33 checks passed

ntucker deleted the claude/project-thread-4fyw11 branch October 5, 2026 23:57

Navigation Menu

Repository navigation

ci: Never cancel benchmark runs on master - #4211

ci: Never cancel benchmark runs on master#4211
ntucker merged 1 commit into
masterfrom
claude/project-thread-4fyw11

ntucker commented Oct 5, 2026 •

edited by cursor Bot

Loading

Uh oh!

changeset-bot Bot commented Oct 5, 2026 •

edited

Loading

Uh oh!

vercel Bot commented Oct 5, 2026 •

edited

Loading

Uh oh!

ntucker commented Oct 5, 2026

Uh oh!

chatgpt-codex-connector Bot commented Oct 5, 2026 •

edited

Loading

Uh oh!

ntucker commented Oct 5, 2026

Uh oh!

github-actions Bot left a comment

Uh oh!

github-actions Bot left a comment

Uh oh!

github-actions Bot left a comment

Uh oh!

Uh oh!

This branch was successfully deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

ntucker commented Oct 5, 2026 • edited by cursor Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

changeset-bot Bot commented Oct 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

⚠️ No Changeset found

Uh oh!

vercel Bot commented Oct 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ntucker commented Oct 5, 2026

Uh oh!

chatgpt-codex-connector Bot commented Oct 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codex Review Summary

Uh oh!

ntucker commented Oct 5, 2026

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Benchmark Spread

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Benchmark React

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Benchmark

Uh oh!

Uh oh!

This branch was successfully deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

ntucker commented Oct 5, 2026 •

edited by cursor Bot

Loading

changeset-bot Bot commented Oct 5, 2026 •

edited

Loading

vercel Bot commented Oct 5, 2026 •

edited

Loading

chatgpt-codex-connector Bot commented Oct 5, 2026 •

edited

Loading