Last week I had to find out some scalability stats of Applications Manager (AM). My main concern was to find out whether the database or the Java application was the bottleneck.

The application to be tested was running on RH Linux and the applications being monitored were setup on a Solaris 5.8 with 150 Virtual IPs. I will not be going into the scalability stats itself of AM here, but more on how I calculated certain performance metrics.

I restricted myself to just 5 performance metrics :

1) CPU Utilization of the server

2) Request Processing rate of the database. Here MySQL

3) Data transfer rate of the database

4) Memory Utilization of the Java Application

5) Response Time of the Application’s Web Client

Now how did I get the above stats? Well, I used another AM to do that for me.

The reporting helped me as it gives the hourly average as graphs too.

For the perf stats, i used various monitors marked in ():

1) CPU Utilization of the server - (Linux Montoring)

2) Request Processing rate of the database. Here MySQL - (MySQL DB Monitoring)

3) Data transfer rate of the database. (MySQL DB monitoring)

4) Memory Utilization of the Java Application - App Manager has a (tomcat)

5) Response Time of the Application’s Web Client - Used (URL Sequence) Monitoring

Hence if you need to test your application’s performance even before going into production, you may want to use Applications Manager - It has a Free Edition too!

AM being more of a data collection and processing application, i did not try out user load testing. But Web 2.0 applications may even want to simulate load on your web application using QEngine and see how ‘N’ simulataneous users impacts the backend and Java application.

Good luck!