HP

HPjmeter 4.1 User's Guide

English
  HPjmeter 4.1 User's Guide   

Chapter 1 Introducing HPjmeter 4.1.00.00

HPjmeter is a performance analysis tool for deployed Java applications. It will help you identify performance problems in your production environment as well as during development.

NOTE: You cannot use HPjmeter to monitor Java applets in a production environment.

HPjmeter helps you diagnose many types of Java application problems that occur only after a product is deployed. The types of problems you can identify include:

  • Memory-retention problems

  • Performance bottlenecks in Java code

  • Improper JVM heap settings

  • Certain application logic errors, such as deadlocks

  • Ineffective or problematic garbage collection

These problems may not be apparent or reproducible before you deploy your application because they depend on unique conditions present only in deployment.

With HPjmeter you can also gain a comprehensive overview of certain states of a running JVM and running applications, including details on memory usage, garbage collection, runtime, and class loading, for example. Using HPjmeter's ability to interact with the Java Management Extensions (JMX) component in the JVM, you can also manipulate operations during a monitoring session to control the state of some logging mechanisms, to gather snapshots of stack traces and memory details, and to force a garbage collection.

Features

With its application and JVM metrics, HPjmeter helps close the loop between developers and operations staff.

HPjmeter has two modes of operation: you can use it to monitor live applications, and you can analyze data collected from applications that have been run using profiling or garbage collection options.

Specifically, HPjmeter provides these monitoring capabilities:

  • Dynamic display of heap size and live objects in the heap

  • Dynamic display of garbage collection events and percentage time spent in garbage collection

  • Memory leak detection alerts with leak rate

  • Java Method HotSpots, which represent CPU usage per method

  • Thread views displaying thread states over time

  • Object allocation percentage by method and by object type

  • Method compilation count in the JVM dynamic compiler

  • Number of classes loaded by the JVM over time

  • Thrown exception statistics

  • Thread deadlock detection

  • Visibility into standard MBean attributes, operations, and notifications (JSR 174) within the Java Virtual Machine, with the ability to trigger operations from HPjmeter and enable or disable notifications.

  • Visibility into user-defined MBeans, with the ability to modify attributes and trigger operations in real time, then monitor the resulting application behavior.

  • Multi-application, multi-node monitoring from a single console

HPjmeter provides these profiling capabilities:

  • Graphic display of profiling data

  • Heuristics on inlining, thrown exceptions, and memory leaks

  • Interactive call graph with call count, or with CPU or clock time, if available

  • Per thread, per thread group, or per process display

  • Allocated and residual objects with object reference graph

  • Comparison capability for performance improvement tracking

HPjmeter provides these metrics for garbage collection (GC) analysis:

  • Details and graphical display of object creation rate, changes in cumulative memory allocation and in heap usage as related to GC events and types, and duration of GC events in the recorded time period.

  • Detailed summaries of GC activity and system resource allocation, along with other system and JVM runtime and version data.

HPjmeter also provides these notification functions:

  • Set notification thresholds for abnormal thread termination, excessive compilation, expected out-of-memory error, and memory leak rates.

  • Set notification thresholds for levels of heap usage, process CPU, and system CPU.

  • Enable or disable notifications.

  • Change notification thresholds in real time to efficiently monitor targeted events