Trust is the only metric that matters
A suite that fails randomly is ignored, and an ignored suite is worse than no suite because it costs money and provides no signal. Everything below exists to protect trust.
Back to Automation BlogCI/CD
Running automation in CI that people trust
Sharding, retries, artifacts and quarantine — the pipeline practices that keep a suite credible.
6 min read8 June 2026Tech Vista Global
Shard, do not shorten
Split the suite across parallel jobs instead of deleting tests to hit a runtime target. Playwright and most modern runners support sharding natively.
npx playwright test --shard=${{ matrix.shard }}/4Retry once, report honestly
One retry absorbs genuine infrastructure noise. Two or three hide real defects. Publish flaky tests as a separate category in the report so nobody mistakes a retry-pass for a pass.
Always upload artifacts
Trace, video and screenshot on first failure turn a five-minute reproduction into a ten-second one. Store them for every failed run, not just for the ones someone remembers to debug.
- trace: 'on-first-retry'
- video: 'retain-on-failure'
- HTML report published as a pipeline artifact
- Quarantine lane for known-unstable specs, reviewed weekly
Turn this into hands-on practice
Book a free demo class and work through these patterns with a mentor.
