This morning I awoke to find my RH server was down.
Here is the message found in log
Sep 10 01:05:12 ns1 kernel: Out of Memory: Killed process 5576 (httpd). and so on...
From what I read about it on the web, linux starts shutting down the biggest memory hog and continues down.
It eventually shutdown the entire system.
I had to reboot because I wasn't even able to ssh in.
The same thing happend about a month ago.
I thought it was caused by a couple cron jobs, so I changed them to run a few hours later. This seemed to solve the problem until in reappeared last night. It always happens around 1:00a.m.
A friend of mine suggested tyring to find out what process is using up the memory by adding a cronjob to perform a ps and keep the output in a log, which I have done.
I am wondering if anyone else has any idea's.

Specs
RH 7.2 w/ latest updates
Dual P-III 933, 512MB Mem, 512MB Swap
Runs Apache 1.3.26, Tomcat 4.0.4, DNS and Postgresql 7.2

The idea with ps is very good. To solve your problem you must know what uses all the memory. Httpd is the first killed program, so it may be the biggest one. Try to observe it (look how much memory it uses from time to time). 1am is a time for cron.daily in most systems. Look into it. But it may also be an DOS attempt.

You can also run "atsar" a sar type app (see freshmeat). Allows for finer grained stats on your apps processes and memory reqs.