I just released a new version of the Azure / OMS Management pack on github.
As some of you know we use it in production to bridge the gap between OMS and SCOM leveraging the superior alerting capabilities of SCOM and the analytics we get with OMS.
With this release I added a couple of monitors to track if an agent is not only sending data or heartbeating (we had this in previous releases), but even if specific data types are returned from the agent. Not all channels and data types are equal, we had cases where agents were reported as healthy, but they didn’t return any perf data point or event in reality. With these monitors we can track this condition and remediate the agent. By default we check for Perf and Security Event data types. Obviously an agent is marked as unhealthy only if it has ever (last 10 days by default) returned data for that specific type.
Hope it helps.