HP

HPjmeter 4.1 User's Guide

English
  HPjmeter 4.1 User's Guide   

Glossary

A

active mode 

An HPjmeter JVM agent that is part of an open console session and is collecting data from the JVM and application and is sending it to a console.


agent 

A module that represents HPjmeter in your environment. HPjmeter provides two types of agents, a node agent and a JVM agent.


B

Batch mode 

An HPjmeter JVM agent that is installed, configured, and is collecting data, but is not connected to a console and is not part of a session. The collected data is stored in a file on the managed node that you can open later from the console.


BCI 

See bytecode instrumentation.


bytecode instrumentation 

Instrumentation of Java bytecode inserts a special, short sequence of bytecode at designated points in the Java classes of an application to facilitate runtime analysis of those classes during monitoring.


C

cached session 

A session created when a running JVM terminates on its own, or when you close an active session. The data in a cached session is available for analysis, but it is no longer connected to a running JVM and is no longer collecting additional data.


call graph 

Displays the capture of “call” relationships between the methods. The nodes of the call graph represent the program methods, while the directed arcs represent calls made from one method to another. In a call graph, the call counts or the timing data are collected for the arcs.


call graph trees 

Call graph data presented in a collapsible and expandable tree hierarchy. See also call graph.


clock time 

The time as measured by an external independent clock.


confidence 

The measure of reliance or assurance that HPjmeter diagnosed the situation correctly.


confidence interval 

The HPjmeter graphical representation of probability that the displayed values are correct. As the number of samples taken increases, the probability increases.


console 

The main window containing the menu and button controls for monitoring live applications and analyzing captured data.


Console filters 

The Session Preferences window allows you to specify the classes to measure when you open a session with a JVM agent. For details, see Controlling Data Collection and Display


CPU time 

Time spent by any central processing unit (CPU) to execute a thread.


CPU virtual times 

Times when a thread is runnable, but not necessarily running. Depending on the actual hardware, operating system, and Java VM, this time may also include I/O time, but generally CPU virtual times are a good approximation of real CPU time consumed by the application.


creation rate 

The calculated rate at which new objects are occupying the heap.


D

dormant mode 

An HPjmeter JVM agent that is installed and configured, but is not part of an open console session and is, therefore, not collecting data from the JVM and applications and is not sending data to a console.


dynamic attach 

In HPjmeter, the ability to find and connect to a running JVM on a machine that has the HPjmeter node agent running, but the JVM was started without using HPjmeter options.


E

eden 

That part of the heap where new and/or short-lived objects (the young generation) are created and used.


ephemeral port number 

a temporary port number assigned from a range of numbers designated for this purpose.


exclusive times 

The accumulated total times used by a method, but not including the time used by the methods that were called from it.


F

filter 

To exclude the measurement of specified classes. HPjmeter provides two methods to control how information is collected and displayed, JVM agent filters and console filters. For details, see Controlling Data Collection and Display


G

garbage collection 

The process of automatically freeing memory by removing objects that are no longer referenced by the program.


H

hot spot 

A section of code, typically a method, with a higher-than-normal level of execution or access requiring high CPU usage.


I

inclusive time 

The accumulated total time used by all invocations of the method and all methods that were called from it directly or indirectly.


inlining 

Replacing a call to a function with an instance of the function; usually done at compile time.


J

Java Management Extensions 

See JMX .


JMX™ 

Java Management Extensions. API for managing and monitoring applications, devices, services, and JVMs. For API specifics, see the Java Community Process web (http://jcp.org/en/home/index): Java Management Extensions (JMX) Specification (JSR3) and JMX Remote Specification (JSR160).


JVM 

Java Virtual Machine


JVM agent 

A module running on each managed node that collects data and sends it to the node agent, which sends it to the console.


JVM agent filters 

The include and exclude options to the JVM agent allow you to filter metrics by class name. For details, see Controlling Data Collection and Display


JVMPI 

JVM Profiler Interface


JVMTI 

JVM Tool Interface


L

lingering objects 

accumulated live objects that clog the heap, creating memory leaks.


lock contention 

The summary time measured from the moment the thread requests a Java lock until the lock is actually granted.


Lock Delay 

The time elapsed between the moment a thread tries to acquire a lock and the time the thread enters the monitor.


M

managed node 

A computer on which the HPjmeter node agent and JVM agent are installed.


memory leak 

HPjmeter uses the long term linear upward trends in the heap size after garbage collections to calculate the likely presence of memory leaks. See How Memory Leak Detection Works


metadata 

Objects needed by the Java interpreter and for reflection in garbage collection.


method 

A function defined in a class. Unless specified otherwise, a method is not static.


method call count 

The number of times each function (method) of the program was called during a run.


method clock time 

The actual time spent in each of the methods of the program.


metric 

In this context, a mapping that associates numerical values with static or dynamic elements such as program functions, variables, classes, objects, types, or threads.


N

node agent 

A process that manages communication between JVM agents and a console. Node Agent


O

old generation 

Objects in the heap that have survived a successive number of garbage collections, and have been moved to the old space in the heap.


P

package 

A group of types. Packages are declared with the package keyword.


R

reduction 

Accumulating data during tracing such that consecutive calls from the same caller to the same callee increase the recorded time value for a call.


S

scavenge 

Incremental, lightweight garbage collection of primarily short-lived objects in order to free space.


scope 

The breadth of view across thread-based profile data. HPjmeter lets you look at a single thread, a group of threads, or the entire application.


service-level objective 

service-level objective. A service-level objective (SLO) partitions a service-level agreement (SLA) into metrics and operational information to enforce and/or monitor the SLA.


session 

An open, live connection between an console and a JVM agent running on a managed node. An open session has associated with it a set of metrics, filters, and alerts that are configured when you open the session.


severity 

The measure of the expected impact of the detected situation on the application stability.


SLO 

See service-level objective.


SSL 

Secure Sockets Layer


stop-the-world time 

Duration of a full garbage collection where processing stops during the collection time.


T

thrashing 

Excessive paging due to a disparity between physical memory actually available to the application and the specified maximum size for memory usage. 


thread spectrum 

When profiling, the distribution of thread lifetime over the possible thread states.


time slice 

Time slice entries represent the life span of the current monitoring session for an application. You can view data throughout the life span of a time slice. 


train 

Optional structure within the heap that is intermediate between the eden space and the old space. The train is a related collection of smaller memory spaces (cars) that have garbage collected incrementally according to the train algorithm.


train algorithm 

Mathematical formula that describes incremental garbage collection of related old objects. The purpose of train garbage collection is to reduce the length of pause due to GC that is visible to users of the software.


V

visualizer 

The pop-up screens where you see graphical or tabular displays of data. Visualizers provide contextual controls based on the data displayed. See Using Visualizer Tool Bars


Y

young generation 

New and/or short-lived objects in the heap. Usually garbage collections are done frequently in the young generation by scavenging the eden space.