Periodic Notification of Sidecar agents in Unknown or Failing State

Original request recieved via support ticket:
" I strongly feel that there needs to be something in place to draw our attention to failing sidecars."

## What?

From original request:
"A notification sent out say once a hour with a list of failing sidecars would save us the trouble of having to periodically stop what we are doing and go make sure all sidecars are functioning as expected."

Additional thoughts:
Perhaps we could create a system event that writes a specific event for each registered sidecar agent that is in an unknown or failing state. Use those events for alerting and visualization/reporting.

## Why?

From the original request: "We potentially could end up with hundreds of sidecars deployed and it could be time consuming to go consuming for failed instances.

I know we could filter on Failed/Unknown from the sidecar administration page but we would need to do it throughout the day to ensure they are all in good health."

## Your Environment

* Graylog Version: unknown
* Elasticsearch Version: uknown
* MongoDB Version: unknown
* Operating System:unknown
* Browser version:unknown