It is important to understand how to deploy the HealthCheck for Azure Agent so that you can obtain the assessment results that you want.
An assessment Agent that is installed on any machine (either physical or virtual) is included in the physical infrastructure list. Those machines will only be moved to a virtual infrastructure list if an Agent is installed on the host machine and indicates that the VM is being managed by its vCenter or Hyper V host server.
If you install an Agent only on the host machine, it will collect system-level data for only that machine. If you install the Agent only on the VM itself, it will collect system application-level data. Installing the Agent on the host machine is appropriate only if you want the host itself included in the scope of the assessment.
This pattern sometimes causes confusion when vCener and HyperV servers appear in the list of physical servers rather than the list of virtual servers. This occurs whenever the Agent is not installed on the host. It is not necessary to install the Agent on the host, as explained above.
Additional Agent Installation Issues
You see a variance between the number of nodes under assessment and the number listed on the assessment report.
Occasionally, the assessment report lists a number of nodes that is fewer than the number of nodes under assessment. Investigations reveal error messages in the assessment report to the effect that the "VM never ran." This results because some servers may spin up for a very brief period, are therefore captured as part of the network map, and yet are not running during critical portions of the assessment period.
The meaning of the "Resource Pool" column in the downloaded spreadsheet report
When you download the spreadsheet from the portal, you see a column labed "Resource Pool"; the column is blank for all servers. You can ignore this column. It will be removed in future versions of HealthCheck for Azure.