12/22/2021, 8:27 AM
Perhaps it would be useful to have a “parallellism group” option for tests, where a set of tests can be grouped, and an option telling how many of those that may run in parallell, as I think that running too many in parallell seems to saturate available resources.
Example, observed locally.
Copy code
09:17:15.53 [ERROR] Completed: Run Pytest - src/python/pants/backend/docker/lint/hadolint/ failed (exit code -15).
Exceeded timeout of 60.0 seconds when executing local process: Run Pytest for src/python/pants/backend/docker/lint/hadolint/
When run on all docker tests at once. Singling out the above test (simply re-running, the others had cached results)
Copy code
09:25:21.35 [INFO] Completed: Run Pytest - src/python/pants/backend/docker/goals/ succeeded.
09:25:40.52 [INFO] Completed: Run Pytest - src/python/pants/backend/docker/lint/hadolint/ succeeded.
It completes just shy of 20s.


12/22/2021, 4:14 PM
maybe… but more generally, if there was more assistance in setting timeouts (such that maybe they weren’t enforced locally and were just warnings) would help with figuring out how to set these.
because there can be lots of reasons for contention
💯 1