Since our ExecDB plans didn't fully work out, please add CRASHED outcome so that we allow maintainers to get notified and inspect their crashed jobs (as long as they manage to generate results.yml). See https://pagure.io/task-abicheck/issue/19 for a concrete use case.
CRASHED
results.yml
ResultsDB supports any outcome you want, on top of the default 'PASSED', 'INFO', 'FAILED', 'NEEDS_INSPECTION' set, you just need to set the ADDITIONAL_RESULT_OUTCOMES config value (in /etc/resultsdb/settings.py) properly. In this particular case: ADDITIONAL_RESULT_OUTCOMES = ['CRASHED'] should be what you want.
'PASSED', 'INFO', 'FAILED', 'NEEDS_INSPECTION'
ADDITIONAL_RESULT_OUTCOMES
/etc/resultsdb/settings.py
ADDITIONAL_RESULT_OUTCOMES = ['CRASHED']
I can set it up in Ansible, if you want me to, but closing this as 'not a bug'.
Metadata Update from @jskladan: - Issue close_status updated to: Invalid - Issue status updated to: Closed (was: Open)
Filed a ticket in the right place: https://pagure.io/taskotron/issue/274
The problem is that resultsdb docs state the 4 outcomes as the only ones supported. That might confuse people. Should we at least add a note that additional outcomes might be available, if a particular instance supports them? Or should the docs reflect the Fedora configuration?
Why not. Not that many people actually bother to read the docs, but it makes sense.
IMO, the 'instance docs' should provide the docs on what outcome are available - the Taskotron docs look like a good place from my perspective.
On top of that, the "root" of the api also shows the available outcomes: https://taskotron.fedoraproject.org/resultsdb_api/api/v2.0/