On Mon, Dec 8, 2014 at 7:57 PM, Jeff Trawick wrote:
> (Now that I wasted the last 59 minutes on this I guess someone will tell me
> an existing way to do the same thing :) )
I don't think so, I was actually just looking for getrusage() usage
the other day and was surprised no hits.
--
Eric Cove
On Mon, Dec 8, 2014 at 4:30 PM, Rose, John B wrote:
> Looking at "top" output we have some occasional http requests that
> generate a wildly disproportionate amount of %cpu usage compared to the
> more frequent %.
>
Some httpd child processes may be handling far more requests than the
others.
Looking at "top" output we have some occasional http requests that generate a
wildly disproportionate amount of %cpu usage compared to the more frequent %.
If you want to tie that process listing to the URL being accessed to see if
there was a pattern of a specific web site or page causing that
because those elements are stored in shared memory, you can't
simply just "add a new member" since it would involve allocating
new shmem and copying stuff over, etc. So what we do is, during
init, see how many members you have configed, and then allow
you to add some "empty slots" for growth. That'