CPUutilization

CPU utilization is a measure of how much of a computer's central processing unit capacity is being used over a given period. It is typically expressed as a percentage of busy time versus idle time within a sampling interval. On single-core systems, 100% utilization indicates the processor is fully occupied. On multi-core systems, utilization can be reported per core or as an aggregate, in which case the total value may exceed 100% when multiple cores are active simultaneously.

Utilization is obtained from the operating system's performance counters. Common methods include reading kernel statistics (for example, /proc/stat on Linux) and using tools such as top, vmstat, mpstat, or sar. In virtualized environments, the concept of CPU ready time reflects the portion of time a virtual CPU waits for physical CPU resources, influencing perceived utilization and performance.

Interpreting utilization requires context. High CPU utilization often signals a CPU-bound workload, but it does not by itself determine performance. Throughput and latency depend on queueing, contention with I/O or memory, and the scheduler's behavior. A system with high utilization can be responsive if workload is well balanced; conversely, aging or blocking tasks can cause high wait times even at moderate utilization. Conversely, low utilization can indicate underutilization or bottlenecks elsewhere such as disk or network I/O.

Uses include capacity planning, performance tuning, and autoscaling decisions. Monitoring trends in CPU utilization, together with load averages, response times, and throughput, helps operators assess when to scale resources or adjust workloads to meet service-level objectives.