Alerts for monitoring central

[liam-j-bennett]

We require alerts on both VM and AM in monitoring central.

AM requires a deadman snitch that can be implemented in a service like Pagerduty (https://www.pagerduty.com/docs/guides/dead-mans-snitch-integration-guide/).

VM requires similar alerts to the Prometheus instances in the satellite instances and suffers from the same problem. (#358)

We could implement a deadman snitch, similar to AM, or set up a more centralised meta-monitoring agent for scraping VM that interacts directly with AM.