6.4

Table Of Contents
Table 11. CPU Metric Group
Metric Description
CPU|CPU contention (%) This metric shows the percentage of time the VMs in the
ESXi hosts are unable to run because they are contending
for access to the physical CPUs. The number shown is the
average number for all VMs. The number will be lower
than the highest number experienced by the VM that is
most impacted by CPU contention.
Use this metric to verify if the host can serve all its VMs
eciently. Low contention means that the VM can access
everything it demands to run smoothly. It means that the
infrastructure is providing good service to the application
team.
When using this metric, ensure that the number is within
your expectation. Look at both the relative number and the
absolute number. Relative means a drastic change in value,
meaning that the ESXi is unable to serve the VMs. Absolute
means that the real value itself is high. Investigate why the
number is high. One factor that impacts this metric is CPU
Power Management. If CPU Power Management clocks
down the CPU speed from 3 GHz to 2 GHz, the reduction
in speed is accounted for because it shows that the VM is
not running at full speed.
This metric is calculated in the following way:
cpu|capacity_contention / (200 * summary|
number_running_vcpus)
CPU|Demand (%) This metric shows the amount of CPU resources a VM
would use if there were no CPU contention or CPU limit.
This metric represents the average active CPU load for the
past ve minutes.
Keep this number below 100% if you set the power
management to maximum.
This metric is calculated in the following way:
( cpu.demandmhz /
cpu.capacity_provisioned)*100 .
vRealize Operations Manager User Guide
46 VMware, Inc.