[ previous ] [ next ] [ threads ]
 From:  <EXT dash Mike dot Bradshaw at nokia dot com>
 To:  <m0n0wall at lists dot m0n0 dot ch>
 Cc:  <rick at scorpion dot co dot th>
 Subject:  RE: [m0n0wall] about the load averages: 2.11, 0.07, 0.05
 Date:  Thu, 8 Apr 2004 08:55:52 +0300
time for my 0.02 monetary units

> Hi Rick,
> Rick Ruggiero wrote:
> > It means, the following:
> > 
> > load averages: 2.11% (Avg. process queue last minute), 0.07% (Avg. 
> > process queue last 5 minutes), 0.05% (avg process queue last 15 
> > minutes).
> > 
> > I don¡¯t have the m0n0wall running properly yet, but generally 
> > uptime is the same on all UNIX platforms.
> I think it was previously established that the three numbers
> corrospond to the average percentage for the last 1, 5 and 15 minutes
> respectivly, but I do find it somewhat illuminating that these numbers
> are related to the process queue.

As part of my job, I install monitoring software for Windows and Unix.
On the things that is normally tracked on a Unix machine is Load Average.

The reason I was told when starting out, was that it gives an easy identification of "how hard the
box is working".

The figures I work with (as a rule of thumb) are; 
Less than 1 : Box to big (too little work to do, Personal workstation not running SETI etc)
1-3 : Normal/usual
4-6 : Working hard (would benefit from increasing size of machine, more CPU, RAM, IO etc)
6 and above : Stop being cruel

Usually this was looked as trends over time. Most of the machines I have set-up on (Solaris, AIX,
Tru64), tend to show a nice horizontal line (with wiggles, and the odd spike).

Constantly high Load averages (varies from place to place, but on average, over 5) meant that
investigation should be undertaken by an experienced/expert engineer/guru, to find the cause/bottle