devops(bidi): publish BiDi reports to a git branch instead of Azure

[Yury Semikhatsky (yury-s)]

Summary

• Corp policy no longer allows uploading reports to mspwblobreport blob storage.
• Publish the per-channel csv reports to the orphan bidi-reports branch instead, where GitHub renders them as tables and the commit history gives day-over-day diffs. The branch is created on the first run, so no manual setup is needed.
• Publishing runs in a separate job after the matrix, so only that job needs contents: write.
• Also add the report to the job summary.

Report urls once it lands:

• https://github.com/microsoft/playwright/blob/bidi-reports/moz-firefox-nightly.csv
• https://raw.githubusercontent.com/microsoft/playwright/bidi-reports/moz-firefox-nightly.csv

The AZURE_BLOB_REPORTS_* repo secrets are now unused and can be deleted.

Fixes #42759

[github-actions]

Test results for "MCP"

1 failed
❌ [firefox] › mcp/idle-timeout.spec.ts:19 › closes the browser after the idle timeout and relaunches it on the next call @mcp-windows-latest-firefox

8589 passed, 1446 skipped

---

Merge workflow run.

[github-actions]

🟢 CI is clear — the one failure is a known pre-existing flake

Hi, I'm the Playwright bot and I triaged the failing CI run here.

The only failure is [firefox] › mcp/idle-timeout.spec.ts:19 on Windows, and the exact same test with the exact same error failed on PR #42718 two days ago. This PR only touches the BiDi workflow and its CSV reporter, so it can't reach the MCP idle-timeout path.

Details

Overall: 1 failed, 8589 passed, 1446 skipped. Nothing in this PR's diff (​.github/workflows/tests_bidi.yml, tests/bidi/csvReporter.ts, tests/bidi/playwright.config.ts) is loaded by the MCP suite — the BiDi config and reporter are only used by the tests_bidi.yml workflow, which isn't the one that failed.

Pre-existing flake / infra

• [firefox] › mcp/idle-timeout.spec.ts:19 › closes the browser after the idle timeout and relaunches it on the next call (mcp-windows-latest-firefox) — in the aggregated CI results this test has failed 2 of 102 runs (2%) on this bot and passed the other 100; it is green on all 12 other MCP bots (0 failures in ~1300 runs). The other failure is run 34915545227 on PR Fix spelling of restoring in credentials docs #42718 (sha d979f021), which has nothing to do with this change.

  Both failures share one signature — Firefox on Windows tears down early and the snapshot comes back undefined:

  - ObjectContaining { "snapshot": StringContaining "Hello, world!" }
  + Object { "snapshot": undefined }
  
  Error: browserBackend.callTool: Target page, context or browser has been closed
  

  The test asserts the browser is relaunched after the idle timer fires, so a slow Firefox launch on a loaded Windows runner races the timeout. A retry should be green.

_{Triaged by the Playwright bot - agent run}

[github-actions]

Test results for "tests 1"

2 failed
❌ [installation tests] › playwright-cdn.spec.ts:43 › playwright cdn failover should work (https://cdn.playwright.dev/dbazure/download/playwright) @package-installations-macos-latest
❌ [installation tests] › playwright-cdn.spec.ts:43 › playwright cdn failover should work (https://playwright.download.prss.microsoft.com/dbazure/download/playwright) @package-installations-macos-latest

7 flaky

⚠️ [chromium-library] › library/video.spec.ts:521 › screencast › should capture static page in persistent context Radoslav Kirilov (@smoke) `@frozen-time-library-chromium-linux`
⚠️ [chromium-library] › library/beforeunload.spec.ts:130 › should support dismissing the dialog multiple times `@chromium-ubuntu-22.04-arm-node20`
⚠️ [chromium-library] › library/video.spec.ts:690 › screencast › should capture full viewport `@chromium-ubuntu-22.04-arm-node20`
⚠️ [chromium-library] › library/video.spec.ts:690 › screencast › should capture full viewport `@realtime-time-library-chromium-linux`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-page] › page/page-emulate-media.spec.ts:144 › should keep reduced motion and color emulation after reload `@firefox-ubuntu-22.04-node20`

51971 passed, 1247 skipped

---

Merge workflow run.

[github-actions]

🟢 CI is clear — both failures are a CDN outage, not this PR

Hi, I'm the Playwright bot and I triaged the failing CI run here.

Both failures are playwright-cdn.spec.ts:43 on package-installations-macos-latest, and they are failing right now on main too — including on commit 8bb2c097, which is this PR's own base. The CDN returns HTTP 400 for the WebKit mac-arm64 build; nothing in this diff touches installation or downloads.

Details

Overall: 2 failed, 7 flaky, 51971 passed, 1247 skipped. The diff is .github/workflows/tests_bidi.yml, tests/bidi/csvReporter.ts and tests/bidi/playwright.config.ts — the BiDi workflow and its CSV reporter. None of that is loaded by the installation tests, which run npx playwright install against the real CDN.

Pre-existing flake / infra

• [installation tests] › playwright-cdn.spec.ts:43 › playwright cdn failover should work (https://cdn.playwright.dev/dbazure/download/playwright) and the https://playwright.download.prss.microsoft.com/dbazure/download/playwright variant — same story, same bot. In the aggregated CI results each has failed 15 of 645 runs on package-installations-macos-latest and passed the other 630; both are green on every other installation bot (0 failures in ~630 runs each on ubuntu/windows).

  Thirteen of those 15 failures landed in the last 48 hours, across unrelated PRs and pushes to main — feat(webkit): roll to r2362 #42736, feat(webkit): roll to r2363 #42737, feat(webkit): roll to r2364 #42748, feat(video): allow recording video at a custom fps #42752, feat(chromium): roll to r1246 #42757, feat(screencast): style action decorations, keep cursor visible #42758, fix(trace-viewer): reject unsafe snapshot tag names #42712, fix(reporter): fail perfetto writes when the output stream errors #42715, plus push runs on bf9d04f2, 24b69326 and 8bb2c097. That last one is the base commit of this PR, so the failure predates the change.

  The signature is server-side, not a timeout:

  -- from url: https://cdn.playwright.dev/dbazure/download/playwright/builds/webkit/2364/webkit-mac-26-arm64.zip
  -- response status code: 400
  Error: Download failed: server returned code 400 body 'GatewayExceptionResponse'.
  

  Chromium downloads fine in the same run; only the WebKit mac-arm64 zip 400s, and the test then exhausts all five failover attempts. Looks like a CDN/edge problem with that one artifact rather than anything a retry will reliably dodge — worth a look independently of this PR.

Also seen

• The 7 flaky entries (video screencast, third-party cookies, beforeunload, emulate-media) all passed on retry and are in areas this PR doesn't touch.

_{Triaged by the Playwright bot - agent run}