$ Linux Command

System

iostat

Report CPU and disk IO statistics.

performancediskcpumonitoringsysstat

Additional Notes

iostat reports CPU utilization and device I/O statistics such as read/write throughput and the number of operations per second. It comes from the sysstat package and is a standard tool for spotting disk bottlenecks.

With no interval it prints since-boot averages; with an interval it prints live updates, which is how you watch current load.

Syntax

iostat [options] [interval [count]]

Parameters

  • options: Flags that choose what to report.
  • interval: Seconds between reports when watching live.
  • count: Number of reports to print, then exit.

Common Options

  • -x: Extended, more detailed device statistics.
  • -d: Show device (disk) statistics only.
  • -p [DEVICE]: Include per-partition statistics.
  • -m: Show rates in megabytes per second.
  • -k: Show rates in kilobytes per second.
  • -c: Show CPU statistics only.
  • -N: Show device mapper names when available.
  • -t: Add a timestamp to each report.
  • -z: Omit devices with no activity during the interval.

Examples

iostat

Show since-boot CPU and device summary.

iostat -x 1

Print detailed device stats every second until stopped.

iostat -xz 2 5

Print five extended reports, two seconds apart, skipping idle devices.

iostat -d sda 1

Watch only the sda device once per second.

iostat -m -p ALL 1 3

Show per-partition throughput in MB/s, three times.

Practical Notes

  • Without an interval, numbers are averages since boot, not current load.
  • Use -x to see await, util, and svctm for real bottleneck clues.
  • High %util near 100% with high await often means the disk is saturated.
  • Install sysstat first if the command is missing; enable its collector for history.