Hi
We have been doing that a few times and it does work. However, as Jorge rightly points out, this is not ideal.
This approach we only took a few times when our server became unusable due to high process load or rapid failure in a process spawning thousands of new processes.
Our discipline has always been to figure out and fix the problems identified and prevent us from having to use this mass terminate approach. We have also attempted to adhere to the best practices as closely as we possibly could when designing timers and processes.