Uninett
Support wants a "still-alive" signal sent regularly to prove that everything is up
- Heartbeat for argus itself
We can do this with a cron-job or something sending a stateless incident at known intervals.
The code won't be in this repo but we need to register the task somehow.
- Heartbeat for glue services
Heartbeat for glue services may be an issue elsewhere in the tracker already, it should be linked here if it exists.
Idea for heartbeat: make it a new event type with least priority incident level (5). Then we could have something that only looks for this event type and makes a standard incident with a higher incident level to alert support about something being down. We should not force glue services to have a heartbeat mechanism which means we probably need it to be configurable whether a glue service has a heartbeat or not. The easiest would probably be a new field on the SourceType table, or in order to make it configurable per source, on Source.