Monitor CVMFS on VMs

[cat-bro]

CVMFS transport endpoints have become disconnected on one of the galaxy worker nodes leading to jobs failing immediately (probably all singularity jobs and jobs that use cvmfs data).

GA needs to be detecting failures of cvmfs on all of the VMs and pinging the alerts channel. I think this was being done pre-pawsey, set up by TC.