Some customers notice redundant/duplicate QOS from CPU monitoring when using the cdm and vmware probes.
This article discusses the advantages/disadvantages.
cdm
Some customers choose to continue to use the cdm probe to monitor VMs is for consistency, e.g., for small-to-medium environments.
This approach allows customers to have Windows monitoring setup instructions are the same regardless of whether it is a physical server or a VM.
Another advantage of having a robot with cdm installed, is the fact that you don’t have to worry about the ‘Vmware Tools’ always being up-to-date, and if the tools are not up-to-date, you could see incorrect monitoring data being passed back to the vmware probe.
It also ensures consistency in the QoS. Another advantage is that customers get a lot more monitoring out of their Windows template than they we can get from just the vmware probe (thanks to the full collection of monitoring/system probes, not just cdm).
One other thing to note is that if you have cdm monitoring CPU utilization on VMs, then you should NOT monitor that using the VMware probe so you can avoid any duplication of QoS.
vmware
The vmware probe fetches the CPU, Disk and Memory from the VMware ESX host not from the machine itself. If you want to use the cdm probe on virtual machines make sure you disable the same monitors in the Virtual machine configuration, so you’re not monitoring the same metrics that the VMware monitoring profile already collects.
Adverse impacts of this redundancy might cause extra overhead from having to install robots on VMs, maintaining robots, and communication/traffic overhead between robots and hubs.