AutoMaxFlaky triage playbook
A repeatable routine for turning flaky scenarios into stable ones.
How to spot flakiness early, classify its cause, and act with the tools AutoMax provides at each step.
Detect
automax insights show -p <slug> ranks scenarios by flakiness score over the last 30 runs; the dashboard flags any that passed only on retry.
Classify
Open the run: heal events point to locator drift; API snapshots with 5xx or timeouts point to the environment; before/after screenshots show timing races.
Contain
Quarantine (insights quarantine <fingerprint>) so gates stay honest while you work; the scenario keeps running and reporting.
Fix
Locator drift: accept the page-object patch proposal. Timing: replace waits with assertions on state. Environment: tag @env: or record a HAR.
Verify
automax run -p <slug> --scenario "<name>" --repeat-each 10 --fail-on-flaky, then lift the quarantine.