When using application performance analysis tools, the first step in analyzing a code problem is to find out what triggers it. This could be a user action, such as clicking on a specific button, or using a certain function, or it could be a more pervasive problem exhibited throughout the application.
To create a reproducible scenario in a production environment:
In a testing or development environment, create a test case (with whatever test tools you may be using) that reproduces the problem. Make the test case as narrow in scope as possible. If you are using Rational® Performance Tester tools, you will be able to use this test while profiling; see Collecting performance data from an application in an automated test environment.
It is important to identify the trigger so you can isolate the problem to a specific part of the application, and limit the amount of performance data collected during the next step. If you profile the entire application, the amount of data collected may be overwhelming.