Visual VM Profiler: Episode 3

Introduction

• Presenter: Shan
• Topic: Exploring Heap Dump with Visual VM Profiler
• Purpose: Identify performance bottlenecks without monitoring tools.
• Context: Response to a subscriber's question about performance issues without monitoring tools.

Overview of Visual VM Profiler

• Open Source Tool: Works on Windows and other OS.
• Previous Videos: Covered thread dump and heap dump.
• Current Episode: Detailed exploration of heap dump.

Importance of Heap Dump Analysis

• Performance Bottlenecks: Identifying areas causing high response times.
• Application Example: Apache JM application running on a server.

Analyzing Heap Dump

• Classes by Instance Size: Key to understanding background operations.
• Process: Navigate through object classes to identify space-occupying objects.

Key Observations

• Objects and Classes:
  • Int Object: Occupies 36.9% of heap with only 1% count.
  • Byte Object: 21.5% count, 28.1% heap space.
  • Java.lang.String: 19.8% count, 5.6% heap space.
• Suggestion: Sort by object size to identify maximum heap space usage.

Navigation and Tools

• Options Available: All objects, dominators, GC roots.
• Types and Packages: Sticky class, JNA global, etc.

Detailed Object Analysis

• Example: Int object references and their implications.
• GC Roots: Importance in Java virtual machine memory management.

Understanding App Context

• Definition: Context or data structure related to GUI main frame.
• Class Analysis: Fully qualified class name for GUI interaction.
• Instance Identifier: Unique identifiers like memory addresses or sequence numbers.

Garbage Collection Roots

• JNI Global: Native code references preventing garbage collection.

Conclusion

• Next Steps: Analysis of java.lang package in the next session.
• Closing: Encouragement to subscribe and engage with feedback.