Feature
Release failure analysis
Diagnose failed Playwright runs in Azure DevOps releases and post a Teams card.
A release monitor watches an Azure DevOps Classic Release definition. When a release environment’s Playwright tests fail, PR Bot stores a failure analysis with pass, fail and flaky counts and a list of findings, each a diagnosed test failure with a confidence level.
Failures that match a known flaky area declared in the repo’s context file are treated as known noise, not a new regression. A redeploy triggers a fresh analysis that is compared with the prior attempt. The result is delivered to MS Teams as a failure card.