Update warc-servers to reboot from time to time

[anjackson]

Just had an playback outage, reported because there was a problem with the warc-server talking to TrackDB (likely because it had been down a few days ago). It had outdated information which was breaking playback.

To avoid this in the short term, the service could be made to automatically restart it's workers every few hours or so. This would mean any such outages would always self-heal eventually.

This would mean updating the deployment configuration for the two warc-server services we run.