Express (Distributed operating systems), v8.0 > Troubleshoot and support > Analyzing application server Java system dumps with the IBM Monitoring and Diagnostic Tools for Java - Dump Analyzer

Tutorial: Writing Dump Analyzer modules for WAS diagnostics

Introduction

This tutorial explains how to write a new analyzer module, starting with a very simple module and progressively enhancing it at each step to add more features, each of which highlights one of the key aspects of the Dump Analyzer APIs.

The example used in this tutorial focuses on an analyzer that examines some data structures that are specific to a WAS. However, with very few exceptions (such as the base class and some of the WebSphere-specific utility analyzers invoked), the principles outlined here can apply to any analyzer whose main function is to examine Java data structures from a dump containing any type of runtime or application; not necessary a WAS runtime.

Please refer to the full API documentation in the reference section of the Dump Analyzer tool documentation for more details of the various classes and methods used in this tutorial.

Table of contents

1. Initial setup to write and test WAS-related analyzers?
2. The Hello World Analyzer
3. A Simple Analyzer to Examine Data Structures
4. Write and Formatting Reports
5. Print Groups of ObjectWrapper Fields
6. Work with J2SE Collections
7. A Custom Wrapper Analyzer
8. Extract Values from Data Structures; Exception Handling
9. Invoke Another Analyzer
10. Generate Observations
11. Invoke DTFJ Functions
12. Generate Reports with a Summary and a Details Section
13. Integrate the Analyzer into the Tool Menu
14. Putting it together: the complete source code for the WASThreadPoolsSample analyzer
15. Putting it together: sample output for the WASThreadPoolsSample analyzer