Challenge
The values in the VMware CPU Usage performance graphs and table values are twice as high as those of the Veeam Monitor graphs.VMware vSphere example:
Veeam Monitor example:
Cause
VMware calculates these values with Hyperthreading Technology enabled.
The VMware vSphere client shows that “Hyperthreading” is in the “Active” state under the “General” section on the Summary tab for the ESX(i) host within the virtual infrastructure:
As a result, the “Advanced” performance page for this host shows total values for all processors but with the Hyperthreading feature enabled
Once all object are selected under the Chart Option window, the CPU Usage is displayed for all virtual CPUs and the ESX(i) host itself. Summing up the average values of each virtual CPU and then dividing by the number of virtual CPUs produces a value two times smaller than the TOTAL CPU Usage value for the host.
Veeam Monitor calculates the total CPU Usage without using VMware’s “Total CPU Usage for entire host with Hyperthreading enabled” option. It takes each CPU's Usage and then performs calculations as described above. Thus, in Veeam Monitor, the Total CPU Usage value is two times smaller than CPU Usage for the entire host on VMware’s graphs.
Solution
No solution required as this is the expected behavior.
More Information
If you have any questions or problems, please contact Veeam Support at http://www.veeam.com/support