Jenkins is still shipping without a systemd unit file

[nuclearsandwich]

This is a local issue to track https://issues.jenkins-ci.org/browse/JENKINS-41218
There is discussion on the issue which includes a systemd unit we could potentially adapt for use with the buildfarm.

One of our other new Jenkins clusters crashed without restarting today and provided this grim awakening.

[nuclearsandwich]

I realize upon re-read that I never fully described why this is a problem.
When the Jenkins java process exits it is not reliably restarted by the daemon wrapper nor is systemd via its initscript wrapper aware that the service is no longer available and in need of a restart.

A native systemd unit is the most expedient way to resolve that issue. See also ros-infrastructure/buildfarm_deployment#162 for a description from the earlier generation of buildfarm deployment.