How Do I Identify Cause of CPU Spikes

are you using prefork and is your maxclients set too high?

prefork yes

maxclient, I figured was great, been running without issues since I moved to this bigger Linode, no issues with traffic it's down not UP right now.

# apache2 -V
Server version: Apache/2.2.14 (Ubuntu)
Server loaded:  APR 1.3.8, APR-Util 1.3.9
Compiled using: APR 1.3.8, APR-Util 1.3.9
Architecture:   32-bit
Server MPM:     Prefork
threaded:     no
forked:     yes (variable process count)

apach2.conf

Timeout 300
KeepAlive On
KeepAliveTimeout 5
 <ifmodule mpm_prefork_module="">StartServers          5
    MinSpareServers       5
    MaxSpareServers      10
    MaxClients          80
    MaxRequestsPerChild   0</ifmodule> 

Keep in mind I don't think the server is configured wrong…

I want to know how to identify the script or web page that is causing the CPU to jump… I do not think this is a max client type issue.

Many thanks

You could add the %D or %T format string to your Apache logs and see if certain requests are taking an inordinate amount of time to complete. Note that reasons other than being CPU-bound could increase those numbers, so further analysis will be required.

Thanks Vance

No luck so far and still dealing with issues since the 6th.. have no clue what happened on the 6th, no changes on my end.

@kutu:

Thanks Vance

No luck so far and still dealing with issues since the 6th.. have no clue what happened on the 6th, no changes on my end.

Try lowering your MaxClients. If your apache instances are taking up a lot of memory (ie lots of installed modules), 80 may be too many even on a 2048.

Have you seen a traffic spike since the 6th? It could be that you just weren't getting up to the MaxClients before that. That's how I got bit the first time with this issue. I was fine with low traffic, then a traffic spike actually caused my server to get up to the MaxClients. It was too high and I was screwed.