Linux Server / Performance, Capacity & Containers
Separate Linux load average from actual CPU saturation
A high load average can reflect runnable work or uninterruptible waits. Inspect CPU activity and blocked tasks before adding processors.
Environment and inputs: Linux server baseline · Ubuntu 24.04 LTS / RHEL 9-style systems. Use the specified distribution, init system and installed tool; service names, privileges and supported options can differ.
Start with this check
Confirm the host, shell and access role. Replace sample names, paths, domains and interfaces with approved targets. These are diagnostic commands; assess policy and workload impact before making changes.
vmstat 1 5Read the result
On a system with procps tools, the first report commonly summarizes activity since boot and later reports describe sampling intervals. Read run-queue, blocked-task and CPU fields together.
Resolve and verify
Correlate sustained runnable demand with CPU idle time and I/O waits during the application slowdown. Address the demonstrated limiting resource and retest under comparable load.
Before changing production
Do not treat load average alone as a CPU percentage or infer a bottleneck from the first cumulative sample.