지금 지원 담당자와 채팅
지원 담당자와 채팅

Foglight for Java EE Technologies 5.9.13 - Application Servers User Guide

Monitoring Application Servers Monitoring Systems Monitoring Servers Monitoring Deployed Applications Monitoring Requests Managing Traces Using Object Tracking to Locate Memory Leaks Monitoring Methods Application Servers Monitor Views
JVM view Method Groups view Request Types view Entity EJBs view Message Driven EJBs view Stateful Session EJBs view Stateless Session EJBs view Deployed Applications view JSPs/Servlets components view Resource Adapters components view Web Applications components view Web Services components view .NET views JBoss Services views Oracle Services views Tomcat Services views WebLogic Services views WebSphere Services views JMX Administration dashboard JMX Explorer dashboard
Appendix: Regular Expressions

Examining object tracking data

The results of any object tracking requests for Foglight for Java EE Technologies are listed on the Object Tracking tab of the Java Virtual Machine (JVM) view.

Figure 16. The Object Tracking tab view.

Review the Live and Expired values. Classes with a large Expired value should be considered for investigation.

The Object Tracking tab contains a table that shows class group data for the selected JVM. This table contains the following information:

 
Table 21. Information contained in the class group data table for a selected JVM.
Data Item
Description

Class

Name of the class.

Package

The package of the tracked class.

Is Directly Tracked

This column is checked if the class is directly tracked. Classes that are directly tracked are those specified in the ObjectTrackingClasses property in agent property file. Classes that are not directly tracked are allocated by directly tracked classes, but these classes are only tracked when allocated by a directly tracked class.

Tracking Per Request Type

If checked, classes are tracked on a per request type basis when object tracking is enabled. It is recommended that per request type tracking is done only when actively investigating a specific class. Narrowing down your investigation places less overhead on your JVM.

Health

State of the class health. You can click the icon in the Health column to obtain alarm information.

Objects > Tracked > Trend

Trend of live objects displayed in a spark line.

Objects > Tracked > Live

The number of live objects tracked.

Objects > Tracked > Expired

The number of tracked expired objects.

Objects > Allocated

The number of allocated objects.

Estimated Size (bytes) > Tracked > Trend

Trend of live objects displayed in a spark line.

Estimated Size (bytes) > Tracked > Live

An estimated size of the tracked live objects.

Estimated Size (bytes) > Tracked > Expired

An estimated size of the tracked expired objects.

Estimated Size (bytes) > Allocated

An estimated size of the allocated objects.

Tracked Lifespan(s) > Expired and Collected

The estimate value of Expired and Collected life spans.

Reviewing track per request type data

The Object Tracking tab also offers the option to track by request type. This option is useful when you have identified a suspected memory leak source.

To enable tracking per request type:
1
On the Object Tracking tab of the Java Virtual Machine (JVM) view, select a class in the Object Tracking table.
2
Click Track Per Request Type.
3
Click Enable.
4
Wait for the data to collect.
5
Click the name of the class (in the Class column) to review a list of request types that allocate the object of the class.
From the Object Tracking Class view, you can review the metrics to identify a list of request types potentially causing the memory leak. Then you can collect single traces for that request type to find out where the memory has been allocated.

Monitoring Methods

The Methods tile on the Application Servers Monitor dashboard allows you to monitor specific named methods. From here you can review the invoked operations used by specific applications.

 
* 
NOTE: Method monitoring is only applicable to Foglight for Java EE Technologies.

After your administrator has configured the named methods to monitor (see Collecting metrics on Java methods ), you can use the Methods view to investigate how often a named method is called and to monitor the results of those called methods. For example, you can use Named Methods to monitor specific methods that connect to external systems.

All calls to the defined methods are tracked. The information presented is independent of request sampling. The metrics capture the behavior of the methods invoked during a request.

You can also find information on the Java request types that invoked the named method. For example, the number of log ins, calls to external system, and so on.

The Methods view shows you detailed information about the named methods being monitored in your environment.

From this view, you can perform the following actions:

•
Monitor metrics associated with named methods.
•
Review the invoked operations used by specific applications.
•
Review charts that display metrics about the named methods.
•
Collect Single Traces (from the action panel General tab)
To view the named methods:
1
On the navigation panel, under Dashboards, click Application Servers > Monitor.
2
On the Application Servers Monitor dashboard, click the Methods tile.

The Methods view displays a Request Types table and corresponding metric chart for the selected method. In the chart area, you can quickly review the Call Count, Execution Time and Exceptional Exit metrics for the selected method.

 
Table 22. Information contained in the Method Group table.
Data Item
Description

Name

Name of the method group.

Method group names at the top level are metrics at the domain level. Method Group names in the sub-items are at the server level.

Health

Health status of the method group.

From this link you can access a list of outstanding Alarms and information on the changes made.

System

Name of the system the method is connected to.

From this link you can select servers, applications, and clusters in your infrastructure that the method is calling.

Calls

Total number of calls for the selected time period.

Execution Time (s)

Execution time, in seconds, for the current period, the period average, and the period total.

Exceptional Exits

Total number of exceptional exits for this method.

Exceptional Exit%

Percentage of invocations that exited in exception.

Requests

Number of requests that reference the selected method.

Collecting metrics on Java methods

The Methods view allows you to view metrics selectively collected on one or more Java methods. To collect data for the Methods view, specify the methods to track in the agent instrumentation configuration. To define these methods you can specify the fully qualified class name, method name, and method arguments in the agent configuration profile.

 
* 
NOTE: It is recommended that you make a backup of the default agent profile before editing.
NOTE: Restart your instrumented application servers after editing the instrumentation configuration (agent/instrumentation.config) in order for the changes to take effect.

For example:

You can group overloaded methods together. In the example below, we are defining a method group which uses similar methods. Get Balance is displayed in the user interface and all the methods within that group are aggregated.

NamedMethods = {
"Get Balance": MethodList(
include "com.globex.Account.close()",
include "com.globex.Account.close(Ljava/lang/Integer;Z[B)I|)”,
include "com.globex.Account.close(Ljava/lang/String;Z[B)I)",
),
}

 

To define named methods:
1
On the navigation panel, under Dashboards, click Application Servers > Administration.
2
On the Application Servers Administration dashboard, click the Java tab.
3
On the Java Administration page, click the Agent Configurations tab.
4
Create a backup of the configuration that you are about to edit.
a
Click the configuration name.
b
In the menu that opens, select Copy.
c
In the dialog box that opens, type a new name (for example, Default_Backup) in the New ID box.
 
* 
NOTE: You cannot use spaces in this name. Underscores, hyphens, periods, and alphanumeric characters are allowed.
d
Click Copy.
e
A message box opens. Click Close.
5
In the list of agent configurations, click the name of the configuration again.
6
Select Edit instrumentation settings.
The Configuration Category Editor opens.
7
Click the Named Methods tab.
8
Click Edit to the right of NamedMethods.
9
In the Settings Editor, click Add named method.
10
In the dialog box that opens, type the name of the named method
11
Click OK.
The Settings Editor refreshes and a table displays on the right.
12
Click Add to add a row to the method list table.
13
Click in the Name box and type the name of the method.
 
* 
TIP: To use a regular expression, click the Regex check box.
14
To add another method, repeat steps 11 and 12.
15
After adding all the methods, click OK.
The Settings Editor closes, and the named method appears in the list on the Named Methods tab of the Configuration Category Editor.
 
* 
IMPORTANT: These settings are valid only for FullDetail and ComponentDetail Instrumentation Levels. Changing the values has no effect if the Instrumentation Level is BasicDetail or NoDetail. You can set the instrumentation level on the General page of the Configuration Category Editor for the agent instrumentation profile.
16
Click Save. The Configuration Category Editor closes.

Understanding the implications of using named methods

Although the NamedMethods MethodList accepts a broad package definition, it is not recommended that you define your NamedMethods this way. Excessive NamedMethod instrumentation may increase the load on the Foglight Management Server and overhead on the application server, so it is best to include only the methods that you are interested in. To guard against excessive instrumentation, there is a limit on the number of NamedMethods that can be instrumented. Once this limit has been reached, no more methods are instrumented to display in the Application Servers Monitor > Methods view. A warning is logged for each method that is configured to be instrumented but is not instrumented when exceeding the allowed number of NamedMethods.

If you have configured too many NamedMethods and you are getting this warning, the first course of action is to modify your NamedMethods MethodList to only include the methods you are interested in or to be more specific in your named method definitions. If you have done that and are still getting warnings, you may increase the limit by editing the MaxNumberOfMethodsTracked property in the agent instrumentation profile. The default value is 100.

 
* 
TIP: You can set the MaxNumberOfMethodsTracked on the Named Methods tab of the Configuration Category Editor.

For more information about editing agent instrumentation configurations, see the Foglight for Java EE Technologies Installation Guide.

 

관련 문서

The document was helpful.

평가 결과 선택

I easily found the information I needed.

평가 결과 선택