PRM (Process Resource Monitor) already does that. My script is designed to kill Apache before it gets to a non-functioning state and eventually crash the complete server. I had a problem with an earlier version of Apache, such behavior occurred frequently. However, an upgrade solved this which means the script is no longer needed. I'm not certain that my experiences with the older version of Apache were isolated incidents. The binary could have been corrupt, for example. You could alter the script and only kill the PID's that are causing the high server load, for more general purposes.