Tests are failure prone
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
kibana (Juju Charms Collection) |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
I plan on re-visiting this bug, pardon the sketch notes, I'll refactor later.
11-scale-elastic.py will only ever work on the first execution. If the test env is configured to not rest for example, the test will fail every subsequent run.
This is problematic, and the test should be refactored to use a polling scenario to ensure the data transmission succeeded and has fully registered in elastic search.
It should then further more evaluate rather than assume that it was the only agent transmitting data, so examine the data set and find a unique ID that was sent to elastic search? (area for improvement)
see linked PR that raises the details: https:/
Related branches
- Review Queue (community): Needs Fixing (automated testing)
- Cory Johns (community): Needs Fixing
-
Diff: 231 lines (+102/-27)9 files modifiedactions/load-dashboard (+14/-2)
config.yaml (+6/-0)
hooks/config-changed (+7/-0)
hooks/install (+9/-4)
tests/00-setup (+0/-16)
tests/10-bundles-test.py (+44/-0)
tests/11-scale-elastic.py (+16/-5)
tests/bundles.yaml (+3/-0)
tests/tests.yaml (+3/-0)
description: | updated |
Changed in kibana (Juju Charms Collection): | |
status: | Fix Committed → Fix Released |
Hi there, here is a patch you can consider for this issue.