Meanwhile we have too many pipelines so it can happen that some checks
wait longer than 30 minutes before they get scheduled - but the timeout
affects this as well. So we need to monitor this
- Drop Leap:15.1:Images and Leap:15.2:MicroOS
- Some .yaml files were changed manually, adjust the .erb source
- Delete some unnecessary/duplicate whitespace
We start most jobs with --debug to generate output, but we had the
case that OBS network caused the connection to just get stuck and
then it 'paused' for 18 hours
This way we don't rely on installed plugin (to avoid the deployment
cycle) and we don't need to install into the real home of the go user
to avoid risking that a broken plugin breaks all future jobs of the
agent
