Playwright trace viewer keeps timing out on CI runs

Question

We run Playwright 1.45 in GitHub Actions with 6 shards. Locally the trace viewer opens fine. On CI, uploading trace.zip fails with a 15 min timeout on ~20% of runs. Anyone hitting the same? Would love a proven artifact upload pattern.

Priya Gopal · Accepted Answer

The 15 min timeout on trace.zip almost always means the upload runner is doing too much. Two fixes we shipped: 1) only capture trace on failure, and 2) upload traces as a separate step with a shorter path list so v4 does not re-hash your whole node_modules. Bonus: use if-no-files-found: ignore to keep green shards from failing.

// playwright.config.ts
use: {
  trace: 'retain-on-failure',
  video: 'retain-on-failure',
  screenshot: 'only-on-failure',
},

# .github/workflows/e2e.yml
- name: Upload traces
  if: always()
  uses: actions/upload-artifact@v4
  with:
    name: playwright-traces-shard-${{ matrix.shard }}
    path: test-results/**/trace.zip
    if-no-files-found: ignore
    retention-days: 7

Mei Liu · Answer

Also worth checking: retention-days lower than 30 dramatically cuts upload time on Azure-hosted runners. And do not zip traces yourself - Playwright already zips them.

Playwright trace viewer keeps timing out on CI runs

2 answers