Quickly measure CPU utilization across cores and spot imbalances. 04.07.2026 | reading time: 2 min When the server feels slow, the administrator needs numbers, not guesses; mpstat delivers a short, tabular snapshot of how each CPU spends time, showing user, system, IO wait and idle percentages in a compact form. Quick hands-on Run `mpstat -P ALL 1 3` to sample every CPU once per second three times; sample output (abbreviated) shows a combined line and per-CPU lines like `12:00:01 all 2.0 0.0 1.0 0.5 0.0 0.1 0.0 0.0 96.4` and `12:00:01 0 0.5 0.0 0.3 0.5 0.0 0.0 0.0 98.7`, where the important columns are the user, system, iowait and idle percentages so the administrator can spot who is consuming CPU time. Fine-tuning the view Specify a processor with `-P 0` or all with `-P ALL` and change sampling by giving an interval and count like `mpstat 2 5`; mpstat reads kernel counters from /proc/stat and reports deltas, so short samples reveal spikes while repeated sampling highlights trends and the "steal" column points to virtualization contention. When it helps most Use mpstat to decide if slow requests are CPU-bound, to detect per-core imbalance, to separate user vs kernel CPU consumption and to identify excessive IO wait or virtualization steal; it is fast, low-overhead and ideal for quick triage before deeper investigation. Complementary utilities mpstat gives per-CPU summaries; pair it with `top` or `htop` to find the offending process, with `iostat` to check disk-induced wait and with `sar` for historical trends so that the administrator can move from a snapshot to root-cause analysis in a few commands. Next move Practice reading mpstat output under load, combine it with live and historical tools, and turn that skill into credentials; consider deepening knowledge with certifications like CompTIA Linux+ or LPIC-1 and an intensive exam preparation at bitsandbytes.academy. Join Bits & Bytes Academy First class LINUX exam preparation. processes utilities troubleshooting Dieser Artikel wurde mithilfe von KI erstellt.