I have a Hyper-V server which is monitored in PRTG, say HSERVER1, and its VMs are all monitored as well, say VM1, VM2.
I have set the dependency at the device level for VM1 and VM2 to be dependent on the Hyper-V service on HSERVER1.
I then paused HSERVER1 for maintenance and the sensors for it all went blue as did the ones for VM1 and VM2 - looks good so far. Then I shut down just the machines VM1 and VM2.
After a short while the various sensors of VM1 and VM2 started to wake up and went red - the DNS being used by the separate PRTG server is actually the two VMs (yes I know, not the best idea).
Why did the sensors of VM1 and VM2 start getting alerted? They should be paused as the Hyper-V is still down from the manual pause of the device.