New to this topic, I have a basic question about profiler like jProbe, etc ---
Suppose I have a
j2ee web application and want to use jProbe to check the bottleneck. Do I have to
1) get the j2ee application start running on my application server
2) get my jprobe installed and up running with the application server
The problem is it is hard to predict when the performance problem appears on the production system. How can I get a "snapshot" of what's the bottleneck while the system is down due to performance problem ? Do I have to have my jProbe ALWAYS up running while my production server is on ? That doesn't make much sense.