Identify flaky test causes and create ~"flaky-test::" labels
To support gitlab-org/quality/engineering-productivity/team#76 (closed) we should begin tracking the common causes to flaky tests. Here are a few quick ones I could think of:
- conflicting state to the test data
- non-waiting selector method for slow elements
- exceed timeout
- incorrect matcher (
eq
instead ofmatch_array
)
This issue is for us to consolidate the common causes with consistent naming, which we can use to gather data either in sisense or triage-ops reports via labels. Obviously we need to slowly identify them as we go, therefore the completeness of the list is not necessary and we can always come back when we identify more.
cc @gl-quality/eng-prod Before opening this issue up to the wider department I'd like to get a sense from you if this is something we want to proceed.
Edited by Jennifer Li