We have a step in our build workflow that opens issues for tests that fail on develop because we can mostly assume they are flaky:
|
- name: Create issues for failed jobs |
While this is true we keep accumulating these issues (currently >100 open) mostly without ever looking at them. Therefore we should probably rework this workflow a bit to make it more easily actionable. Also if we move to merge queues this workflow will stop working altogether.
We also have this flaky test workflow that I am not yet sure what it does: https://github.com/getsentry/sentry-javascript/blob/develop/.github/workflows/flaky-test-detector.yml
Backend team uses junior to get a list of the most flaky tests each week. Maybe we can do something similar and then make this part of triage or something.
We have a step in our build workflow that opens issues for tests that fail on develop because we can mostly assume they are flaky:
sentry-javascript/.github/workflows/build.yml
Line 1276 in b7227bb
While this is true we keep accumulating these issues (currently >100 open) mostly without ever looking at them. Therefore we should probably rework this workflow a bit to make it more easily actionable. Also if we move to merge queues this workflow will stop working altogether.
We also have this flaky test workflow that I am not yet sure what it does: https://github.com/getsentry/sentry-javascript/blob/develop/.github/workflows/flaky-test-detector.yml
Backend team uses junior to get a list of the most flaky tests each week. Maybe we can do something similar and then make this part of triage or something.