when you run the top command, you see load average x.xx, y.yy, z.zz

how does the server measure load average, since the scale goes above 1.00 it cant be %.

Can anyone explain it to me?

and what is normal for a webserver that is quite busy?




Thanx :confused:

The load average (3 are shown) are the average number of processes waiting for the attention of the CPU over a period of time.

The period of times of the 3 shown are 1 5 and 15 minutes.

Therefore, the first figure is the average number of computational processes over the last minute. The second number os the average over the last 5 minutes and the final figure the average over the last 15 minutes.

If you want a perfectly running server with no CPU contention then you want the figure to be <1 average of any extended period of time.

Oh, as to your final question, anything between 0 and 2 is probably OK, anything over that and you have a resource bottleneck somewhere (could be CPU power, memory, disk IO, etc).

okay thanx for the fast answers!

the server I am on does peak at 1.5-2 over the last 1 minute

but the 15 minute period its always below 1

I wouldn't worry about it, then :)