Auto quarantine spec that has spiked count within a short period of time
Originally discussed in gitlab-org/gitlab#450198 (comment 1881376408)
We should expand our automated quarantine to also target the failures that have spiked within a short period of time.
For example:
In gitlab-org/gitlab#463147, the spec was created and promoted to flakiness2 within a 8 hour window - this signals a spike of failures suggesting high severity of flakiness, thus we should treat it as flakiness1 and quarantine right away.
Suggestion
Identify a threshold for duration and count - if occurrences has exceeded the threshold count after the threshold duration, we immediately change the label to flakiness1;
Edited by Jennifer Li