Answer:

The different data sets are expected due to the different collections methods for virtual machines versus operating system (OS) level polling.

When viewing data on the Device Details page, the data presented will default to what the VMware host is reporting for device. The Economics Report will default to OS level polling. We have found that the VM level polling is more accurate.

This can be verified by selecting that device on each collection method in the Private Cloud Resource Requirements and selecting Committed/Active as the display counter

The collection methods can have different values as these values are reported from the OS level, as well as the VMware level.

The platform uses SSH, SNMP, WMI, as well as the VMware API to collect data if both OS and VMware credentials are provided. We have seen cases where resource utilization can be reported differently from the OS level than what is reported from ESXi/vCenter.

Please note that VMware automatically collects the performance/inventory data at the hypervisor level. The platform queries this information and this information is provided in the platform. The platform also collects metrics directly from the OS level (WMI, SSH, etc) and then calculates this data for performance metrics.

Because Windows and Linux OS's are polled on 5 minute intervals, resource numbers can differ as VMware metrics are polled more frequently. We commonly see VMware metrics more accurate as the polling interval is more frequent.

Also, at this time, there is no feature to change KB to GB or MB other than to convert the numbers after exporting the data into a spreadsheet.