Today one of the compiler workers ran out of disk space, and the only indicator was failed job output.
To address this I can think of a few actions:
- Schedule a cleanup job to remove old directories from the jenkins workspace
- Add disk capacity to the compiler workers (which afaik in openstack means adding cpu/ram resources as well)
- Alert on PCC worker node filesystem near full
- Have the CI job or puppet compiler to compile the huge .pson catalog