Understanding CPU Usage (%) in Aria Operations/VMware environments

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

The Metrics exceed the 100 % utilization
CPU Usage (%) is often misunderstood in performance metrics, Many think it shows how “fast” the CPU is working or that a high value means the CPU is overloaded. This can lead to wrong conclusions when checking performance.

Environment

Aria Operations 

VCF Operations 9.x

 

Cause

The CPU Usage (%) metric measures the percentage of time a CPU was active during the sample period. It does not represent:

• How fast the CPU executed instructions

• The actual throughput of workloads

For example, if the CPU is active 9 minutes out of 10, usage will show 90%. This only means the CPU was busy for 90% of the time. It does not always mean the CPU is running at full capacity.

Resolution

To correctly check CPU performance:

1. Understand that CPU Usage (%) is time-based, not throughput-based.

2. Use other metrics:

   • CPU Ready (%): This metric indicates the percentage of time in which the VM was waiting in line to use the CPU on the host.

   • Co-Stop (%): Percentage of time the VM is ready to run, but is unable to due to co-scheduling constraints.

   • CPU Contention: CPU contention as a percentage of 20-second collection interval.

3. Avoid relying solely on CPU Usage (%) when diagnosing performance issues.

4. Combine usage data with latency and contention metrics for accurate analysis.
   
   

VM Metrics Doc for reference: Virtual Machine Metrics