ci: shard nightly browser projects

[0xkkonrad]

Why

The ordinary browser nightly currently runs both Playwright projects through one Next dev server with six workers. On run 31454365199 that meant 302 tests, 71 failures, 27 flaky tests, and 43.3 minutes. Run 31144205766 showed the same collapse on an older mainline SHA: 81 failures, 20 flaky tests, and 43.0 minutes. Mobile and desktop contributed roughly equal failure blocks, dominated by 30-second timeouts and closed-browser errors.

What

• Split the ordinary matrix by browser and Playwright project, producing one mobile and one desktop job per engine.
• Give every shard its own hosted runner, Postgres service, and Next dev server.
• Run three workers per server, preserving six aggregate workers across the two project shards without saturating either server.
• Give reports project-qualified artifact names.
• Validate every shard plan before browser installation; it must select a non-empty set from exactly one known project.

No product code or browser assertion changes are included.

Local proof

• mobile plan: 151 tests in 31 files
• desktop plan: 151 tests in 31 files
• Chromium mobile command smoke: 3/3
• Chromium desktop command smoke: 3/3
• root typecheck
• Prettier and Node syntax
• Actionlint 1.7.7, excluding the pre-existing SC2034 warning at nightly-deep.yml:113
• git diff check

Hosted proof required before ready

Force Nightly deep verification on this branch and compare all six ordinary shard durations, retry counts, and conclusions with runs 31454365199 and 31144205766. The PR stays draft until the hosted-runner result establishes whether three workers per isolated server is stable.

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

---

Note

🎁 Summarized by CodeRabbit Free

Your organization has reached its limit of developer seats. For new users, CodeRabbit will generate a high-level summary and a walkthrough for each pull request. For a comprehensive line-by-line review, please add seats to your subscription by visiting https://app.coderabbit.ai/settings/billing. If you believe this is a mistake and have available seats, please assign one to the pull request author through the team management page at https://app.coderabbit.ai/settings/team-management.

_{Comment @coderabbitai help to get the list of available commands.}