How to measure CPU usage and performance using vmstat ?

vmstat reports information about processes, memory, paging, block IO, traps, and CPU activity.

vmstat [ -n ] [ delay [count] ]

• -n : causes the header to be displayed only once rather than periodically.
• delay : delay between updates in seconds. If no delay is specified, only one report is printed with the average values since boot.
• count : number of updates. If no count is specified and delay is defined, count defaults to infinity.

vmstat [ -V ]

• -V : results in displaying version information.

Example

vmstat 1 500

Field description for vm mode

Process

• r : The number of processes waiting for run time.
• b : The number of processes in uninterruptible sleep.

Memory

• swpd : the amount of virtual memory used.
• free : the amount of idle memory.
• buff : the amount of memory used as buffers.
• cache : the amount of memory used as cache.
• inact : the amount of inactive memory. (-a option)
• active : the amount of active memory. (-a option)

Swap

• si : Amount of memory swapped in from disk (/s).
• so : Amount of memory swapped to disk (/s).

I/O

• bi : Blocks received from a block device (blocks/s).
• bo : Blocks sent to a block device (blocks/s).

System

• in : The number of interrupts per second, including the clock.
• cs : The number of context switches per second.

CPU

These are percentages of total CPU time.

• us : Time spent running non-kernel code. (user time, including nice time)
• sy : Time spent running kernel code. (system time)
• id : Time spent idle. Prior to Linux 2.5.41, this includes IO-wait time.
• wa : Time spent waiting for IO. Prior to Linux 2.5.41, shown as zero.