Sunday, 5 July 2009

atop - AT Computing's top

Last month I blogged about htop, a process monitoring tool. Today, while I was searching for a way of monitoring disk activity, I stumbled across atop (AT Computing's top). While it shares a lot of similarity with top, it is distinct enough to be useful in it's own rights. Because it is an interactive text based tool, it is driven by simple key presses, for example, pressing 'h' brings up a help page showing how to drive the tool.

Some useful key commands are as follows:

g - generic info (default)
m - memory details
d - disk details
n - network details
s - scheduling and thread-group info
v - various info (ppid, user/group, date/time, status, exitcode)
c - full command-line per process name but a few.

One feature I especially like is atop highlights in red colour the processes that may be misbehaving, such as CPU, memory or disk I/O hoggers; very helpful when a rogue program is killing your machine or sucking away power or I/O.

To install atop, simply use:

sudo apt-get install atop

Below are some screen shots:

atop highlighting CPU hogging

atop highlighting I/O hogging

So, while I may always turn to ps,top or vmstat as my first port of call when I want to check my systems activity, atop is another very useful utility in my toolkit that comes in useful in seeking and resolving misbehaving systems. Check it out and let me know what you think!


  1. There is also a kernel patch to add per-process stats, which it looks like Ubuntu have already added :)

    Without this it can give you general stats, but the per-process accounting extensions it gives for disk and network I/O are exceptionally helpful!

  2. Thanks Legooolas for that observation. :-)