Why does the cpuUtilisation value reported in Hardware plugin appears to be high on Windows Netprobes?

For Windows servers with many CPUs, the Hardware plugin could be sensitive to report high usage because of this design mechanism. In this case, you may consider to use the CPU plugin rather than Hardware plugin for monitoring CPU usage.

Please note that the current behavior for Unix type Netprobes is to report the average value across all CPUs, which would be 25% in the example. This is, however, inconsistent with the documentation.

P.S. There are two change requests under consideration to provide more consistent behaviour for Windows and Unix type Netprobes: PLI-1818, PLI-1820.