[tlug] load averge question

[Jean-Christian Imbeault <jc@example.com>]

I have set up a public web server (apache) on a RH 7.3 machine and the 
load average on the machine goes spikes up pretty high (>3) at the 
oddest times.

I wrote a little script to capture the load average at 5 second 
intervals and I get this kind of spike:

load average: 0.17, 0.17, 0.17
[gradual increase ... ]
[2 minutes 10 seconds later ..]
load average: 6.45, 2.77, 1.15
load average: 7.30, 3.01, 1.23
load average: 6.79, 2.97, 1.23
load average: 6.25, 2.92, 1.23
load average: 6.23, 2.97, 1.25
load average: 5.73, 2.92, 1.24
load average: 5.83, 2.99, 1.27
load average: 5.36, 2.94, 1.27
load average: 5.89, 3.09, 1.33
load average: 5.58, 3.07, 1.33
load average: 6.02, 3.20, 1.38
load average: 5.93, 3.23, 1.40
load average: 5.46, 3.18, 1.39
load average: 5.42, 3.21, 1.41
load average: 6.43, 3.45, 1.50
load average: 6.71, 3.56, 1.54
load average: 6.50, 3.57, 1.56
load average: 7.10, 3.74, 1.62
load average: 7.57, 3.90, 1.69
load average: 7.60, 3.96, 1.72
load average: 7.31, 3.96, 1.73
load average: 7.21, 4.00, 1.75
[goes slowly back down ... ]
[3 minutes 20 secs later back to normal ...]
load average: 0.38, 2.36, 1.65


I'd like to figure out just why this is happening. How can I pinpoint 
the cause of the load average spike? Something automated would be nice 
since I don't know when the spike will happen.

Could the spike be cause by Apache spawning a bunch of server to meet 
it's MinSpare clients setting?

Any hints welcomed!

Jc