Details
Description
When Jenkins jobs are killed due to exceeding their built-in timeouts, useful artifacts that would help with debugging the problem are often not collected and made available via the UI. Arrange for all artifacts that exist and that would be collected in the event of normal termination to be collected in the event of timeout (especially log files).
The jobs 22-24 in the test setup https://ci.lsst.codes/blue/organizations/jenkins/debug-stack-os-matrix/activity
have different timeouts on the tasks: 20m, 40m and no timeout to simulate a build task running into its timeout.
status.yaml is not present in that case, but the build artifacts are uploaded anyway without using the package information. pytest XMLs are uploaded but coverage XMLs are excluded.