Labels

master-broken incidents caused by a bug in the CI configuration

master-broken incident caused by dependency upgrades to Ruby version, Node version, gem versions, etc..

This is to indicate master broken incidents caused by external dependency issues such as npm registry down, RubyGems.org down, etc.

master-broken incidents caused by flaky tests

master-broken incidents occurred in fork repos, main FOSS and JiHu, due to code changes to the Gitlab main repo that does not have any test coverage for running in a fork repo.

This label is for master broken incidents occured because gitaly is failed to setup. See example: gitlab-org/quality/engineering-productivity/master-broken-incidents#625

master-broken incidents caused by various infrastructure issue such as network error, runner memory limit reached, etc.

This label represents a broken master pipeline caused by failing to pull docker image from the image registry.

This label describes a broken master scenario with the error "Gitlab is unable to handle this request due to load."

master broken incidents caused by a gitlab outage or service disruption. Retry is not recommended for such incident unless we are certain the issue is fixed.

When master pipeline incident occurs due to the runner having no disk space left.

This label is to identify master broken incidents caused by job timing out

Broken master pipelines caused by bugs introduced to the repo due to missing test coverage

Master broken incidents caused by db:migrate:multi-version-upgrade job, see example: gitlab-org/quality/engineering-productivity/master-broken-incidents#5982

master-broken incidents caused by merging two merge requests impacting the outcomes of the same spec, potentially resulted in delayed pipeline feedback, would have been avoided if merge train is enabled

master-broken incidents caused by MR being merged without running a new pipeline

Test data state leaked through test cases. The same as ~"flaky-test::state leak" but it's considered broken because it's almost always happening. See https://docs.gitlab.com/ee/development/testing_guide/flaky_tests.html#state-leak

master-broken incidents caused by gap in the test selection process, meaning the tests that would have caught the issue was mistakenly filtered out via MR pipeline test selection.

The cause of the failure is not straight forward and yet to be determined

Used to track broken masters and their fixes.