Split Playwright CI runs?

Created on 25 August 2025, 3 days ago

Overview

I threw the "?" in the issue title in case there's a better way to address.

On CI, I can re-run Playwright multiple times and have different tests fail each time. I figure if they are split into smaller groups, each smaller group is more likely to get a green with fewer retries. Perhaps an irritating idealist might say "how about just get the tests running better" but come on this stuff is tricky and there's good folks working on that already. This would make things easier in the meantime.

Proposed resolution

User interface changes

πŸ“Œ Task
Status

Active

Version

1.0

Component

… to be triaged

Created by

πŸ‡ΊπŸ‡ΈUnited States bnjmnm Ann Arbor, MI

Live updates comments and jobs are added and updated live.
Sign in to follow issues

Comments & Activities

  • Issue created by @bnjmnm
  • πŸ‡ΊπŸ‡ΈUnited States bnjmnm Ann Arbor, MI
  • πŸ‡ΊπŸ‡ΈUnited States mglaman WI, USA

    I've had better luck with Playwright by using their sharding over multiple jobs. We have multiple jobs and pass `--shard=1/3` or `--shard=2/3`, etc to each. Then Playwright knows how to chunk the tests.

    We've also used some tagging for specific tests as well, like `--grep=@multilingual`. But then other jobs need `--grep-invert=@multilingual`

  • πŸ‡¬πŸ‡§United Kingdom justafish London, UK

    we first need to reformat some of the Playwright tests with the new package and split them into ones which can run in parallel and ones which need to run in isolation, which should stablise them 🀞

Production build 0.71.5 2024