Breakages found in under five minutes
- Company
- LaunchGood
- Role
- QA Automation Engineer
- Year
- 2025
- Stack
- Playwright
- GitHub Actions
- Slack
- Result
- Detection from hours to under five minutes
The problem
Users found breakages in production before the team did, often hours after they started.
What I did
- Built the company's first Playwright end-to-end suite from scratch, with reusable page helpers other engineers extended.
- Ran it every five minutes against production on a schedule.
- Sent a Slack alert the moment a check failed.
The key decision
Run the suite on a timer against production, not only on code changes, so it catches breakages from any cause: a deploy, a third party, or bad data.
The result
Breakage detection went from hours to under five minutes.
What I'd do next
Group alerts by user journey, so one outage sends one clear alert.