Identify flaky specs/infrastructure issues by re-running pipelines with green master
From discussions in 2019-12-10 Engineering Productivity meeting, Engineering Productivity should investigate creating an automated way to identify flaky specs and infrastructure failures by re-running a green master commit over and over and triage failures.
Proposal
- Identify green master commit
- Run pipeline multiple times at that SHA
- Identify and triage failures
Benefit
By re-running a known-green any failures would be an indication of flaky specs or infrastructure issues.