Speaker: Neelima Bawa
Troubleshooting production Java applications is a challenging task. Most of the monitoring and diagnostics tools are inadequate in a production environment because of one or more of the following reasons:
" Requires code changes due to byte code instrumentation or AOP techniques
" Requires server restarts due to application changes
" Very high overhead to get enough granularity required for triaging performance problems
" Unable to identify and resolve memory leak issues in a production environment
" Provide no visibility from Java EE containers through Database
In this presentation, Neelima presents a new approach to diagnose production applications by peeking into memory structures of the Java Virtual Machine. She will showcase this new technique by demonstrating Oracle's AD4J product and how it provides the ability to view the state and execution context of application in the JVM with little to no overhead. You will see that this approach does not need any complex configuration or application instrumentation. You will learn to use this methodology to diagnose problems in real time on a production environment, without requiring server restart or application rewrite.
