Skip to content

Monitoring & Understanding Server Load

I have a site whose performance could stand some improvement.  I want to be able to make sure that any changes I make are actually having an impact, so I’m going to start with setting up some monitoring tools. I found a very nice and free monitoring application called Load Average. After I got that all set I did a bit of research so I can fully understand and make the most out of the numbers I see in there.


First off, I needed to get a handle on how CPU Load is measured. If you run “cat /proc/loadavg” on a Linux server you will get a string similar to this:

0.02 0.03 0.00 1/437 21084

The first three numbers are the average CPU Load over 1 minutes, 5 minutes, and 15 minutes.

The lower the numbers the better, and the highest it should be is equal to the number of cores you have.  With one core it would be bad if it went over 1, but if you had 8 cores then a value of 1 wouldn’t be a problem, but 8.5 would.  To find out how many cores you have you can run “nproc.”


Check how much memory you’re using by running “free -m.” You’ll get a readout like this:

             total       used       free     shared    buffers     cached
Mem:         11909      10785       1124          1        234       9372
-/+ buffers/cache:       1178      10731
Swap:            0          0          0

The number you want to watch out for here is the used buffer/cache, here it’s “1178.”  That’s the memory that’s being used by the applications currently running on your server.  That number should be less then the total memory + swap memory, here’s that’s “11909.”



Published inTime Wasters

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *