The first threshold is already configured in the default threshold and watches location.
We need a second threshold need to be configured using a watch.
When using watch there is no attribute to get aggregated cpu alarms and only giving single instance alarms.
How do we configure the watcher to give aggregated alarms?
Release : All
Component : Spectrum Core / SpectroSERVER
For a watch, you can compute it in the expression.
For example, if you know there are 4 CPU's on the device and they are instanced 1, 2, 3, 4, your expression could be the following:
((NRM_DeviceCPUUtilization.1 + NRM_DeviceCPUUtilization.2 + NRM_DeviceCPUUtilization.3 + NRM_DeviceCPUUtilization.4) / 4 )
There are issues with this though:
1. Not all devices have 4 CPU's. Some have fewer and some have more
2. The instances may not all be 1, 2, 3, 4.....
Example