Playwright suites that stay reliable

Situation

A release at Uvion moves only as fast as the browser tests people trust. A flaky failure looks like a product bug, and a suite that cries wolf gets ignored.

What I built

I built end-to-end suites with Playwright and pytest. They follow a page-object structure, run in parallel, and publish a report after each run. I also write behavior-driven scenarios so a test states the behavior in language a developer can read without opening the automation code.

How it ran

Playwright runs from Jenkins with the API and contract checks. A real part of the work is flakiness: finding the tests that fail for reasons other than the product, fixing them, and keeping the report specific enough to act on.

What changed

The browser pass is parallel, reported, and part of the release cycle. Developers and QA read the same scenarios when a requirement changes.

What I would change

I would keep a short log of the flakes that were removed. Stability is a result, and it should be as visible as a red build.