Knowing the performance metrics for a system is necessary for identifying resource contention (bottle-neck) and planning for future capacity. Long term monitoring via SNMP and/or tools like Cacti & Nagios are very useful.

Performance Overview

The tools htop and dstat both provide very excellent views of system information.

htop

A screenshot from htop showing colour coded CPU and memory information and a process table.

dstat

A screenshot from dstat showing pretty output, snapped every second.

Here dstat shows averages collected over 30 second intervals.

nstats

Prints statistics (packet/byte count and average packet size) about ethernet network traffic. This includes protocol breakdown on several layers and statistics about TOS, ECN, ICMP code/type and TCP options usage. nstats has full support for IPv6 including ICMP6.