workers should more gracefully deal with absent server services

[h01ger]

when rebuilderd-worker tries to connect to rebuilderd and rebuilderd for whatever reason (eg reboot, or shutdown to do backups, or temporary network problems) is not there, starting rebuilderd-worker fails. and then systemd tries the same again, 5 times, and then marks the service as failed.

and then 5min later rebuilderd is back and the worker could be started, but this has to be done manually. and thats not very clever and annoying.

[h01ger]

the solution might be in the rebuilderd-worker.service file, not sure. i just dont think the solution should be external, be it relying on something like monit or manually restarting the worker once rebuilderd is available again