I need to develop a perl/shell script using linux commands only for finding out how much CPU, Memory and other stats for JVMs on our WAS servers. Problem is when using `top` the JVM keeps forking out many different process and whose cpu keeps spiking so making it difficult to monitor. Do I need to add cpu usage/memory child process of JVMs PID and
I need to whats the best way to tackle this problem?
For Linux, one consideration is procfs. You can get live information, just by reading files. There are a bunch of files in /proc that gives you the state of the system. And a bunch more in /proc/PID that will give you the state of a particular process.