Monitor service crashes when GitHub returns HTTP 503

[delan]

When this happens, the service needs to be restarted manually.

Nov 15 14:05:09 ci0 monitor-start[2405852]: [2024-11-15T14:05:09Z INFO  monitor::profile] Destroying runner 3493 with base vm name servo-windows10
Nov 15 14:05:09 ci0 monitor-start[2540097]: Domain 'ci-servo-windows10.3493' destroyed
Nov 15 14:05:09 ci0 monitor-start[2540136]: Domain 'ci-servo-windows10.3493' has been undefined
Nov 15 14:05:10 ci0 monitor-start[2540145]: will destroy tank/ci/servo-windows10.3493
Nov 15 14:05:15 ci0 monitor-start[2405852]: [2024-11-15T14:05:15Z INFO  monitor] Did not receive an API request
Nov 15 14:05:16 ci0 monitor-start[2540153]: gh: Server Error (HTTP 503)
Nov 15 14:05:16 ci0 monitor-start[2405852]: [2024-11-15T14:05:16Z ERROR monitor] Monitor thread error: Command exited with status exit status: 1
Nov 15 14:05:16 ci0 systemd[1]: monitor.service: Main process exited, code=exited, status=1/FAILURE
Nov 15 14:05:16 ci0 systemd[1]: monitor.service: Failed with result 'exit-code'.
Nov 15 14:05:16 ci0 systemd[1]: monitor.service: Consumed 4min 55.506s CPU time, 47.8M memory peak, 4K read from disk, 7.6M incoming IP traffic, 2.4M outgoing IP traffic.

• Restart systemd services on failure (55b3f42)
• Restart systemd services after extended failures (5 times in 10 seconds) with an hourly timer
• Don’t crash when GitHub API requests fail