Performance bottleneck indicators at different Application layers

When we execute a performance test against an application there are different performance bottlenecks specific to the different layers of the application.

Following diagram show what all issues we can see at any specific layer in a three tier application architecture.

What is the root cause for these issues? We may need to drill down further and find that.

There may be some indicators which may appear at any layer but may be because of issues further down the execution path. For example “High time out errors” at web server layer may be because of slow responding service call to the application layer. Further there may be some request which may need services of any slow responding external system and this also may cause timeout errors. We should apply our judgment and logic while trying to fix responsibilities for the issue at any layer. I have used Introscope and dynaTrace for analyzing these performance issues and those are very helpful in diagnosing these.

How can we drill down to the root cause of these issues? I will try to explain that in next post.