Detect failing nextcloud postStart in CI without monitoring
We recently changed nextcloud postStart so it doesn't crash the nextcloud pod if something fails there.
We may want to adapt the CI so it checks for those errors explicitly, because a minor change to nextcloud -- for example an update to a nextcloud app -- will not trigger the monitoring app to be installed, so the prometheus alert for that error will not fire. We could add a small check to the nextcloud CI test to parse the contents of the postStart status
file that's created in /var/www/tmp
in the nextcloud pod (also available on the node).