7.2.2 Overview of the trace based performance analysis of applications
The trace based performance analysis of applications is a functionality that analyzes the application processing performance by using the performance analysis information (trace information) output when the application processing to be traced is executed in the processes from the start up to the termination of an application (hereafter, the trace based performance analysis of applications is referred to as the user-extended trace based performance analysis). The user-extended trace based performance analysis presumes the use of the trace based performance analysis of the Application Server. This analysis enables you to identify the processing bottlenecks and to improve the efficiency of troubleshooting by checking the extent to which the processing has been attained when an error occurs. Furthermore, you specify the processing for which you want to acquire the performance analysis information in the configuration file for the user-extended trace based performance analysis. You can efficiently check and compare the application performance because the application need not acquire the performance analysis information.
- Organization of this subsection
(1) Collecting the trace information of the user-extended trace based performance analysis
The trace information of the user-extended trace based performance analysis collects the performance analysis information that is output when the processing specified in the configuration file for the user-extended trace based performance analysis is executed.
The following figure gives an overview of the trace information collection for the user-extended trace based performance analysis.
|
When you start the J2EE server, the configuration file for the user-extended trace based performance analysis is read. When the processing (trace collection point) specified in the configuration file for the user-extended trace based performance analysis is invoked, the trace information is output to a buffer. When a certain amount of information is collected, the performance tracer (PRF daemon) outputs the information to a trace file (PRF trace file).
When you use Management Server for operations, you can collect the trace based performance analysis file by editing the PRF trace file in the text format. On the basis of the collected trace based performance analysis file, the operations administrator can perform the entire performance and bottleneck analysis in the management domain. For details on how to collect the trace based performance analysis file and the information that is output, see 7.3 Collecting the trace based performance analysis file by using Management Server.
(2) Working of the user-extended trace based performance analysis
In the user-extended trace based performance analysis, you specify the names of the methods that will be used to acquire the trace information, in the configuration file for the user-extended trace based performance analysis. The following figure gives an overview of trace output for the specified methods.
|
If you enable the user-extended trace based performance analysis, the user-extended trace based performance analysis reads the configuration file for the user-extended trace based performance analysis. If the application is executed and if the methods (methods to be traced) specified in the configuration file for the user-extended trace based performance analysis are invoked, the trace information is output to the following locations:
-
Method entry
Trace information immediately after the method is started.
-
Normal exit of method
Trace information just before the method terminates normally.
-
Abnormal exit of method
Trace information immediately after an exception or error occurs in the method. However, the exceptions or errors thrown at the method invocation source are excluded.
(3) Configuration of the user-extended trace based performance analysis
The user-extended trace based performance analysis is configured from the following elements.
Note that the user-extended trace based performance analysis uses the class load hook processing of the instrumentation functionality and rewrites the applications to be traced in order to output the trace based performance analysis.
-
Configuration file for the user-extended trace based performance analysis
You set up the information about the methods to be traced by the user-extended trace based performance analysis by using the configuration file for the user-extended trace based performance analysis. For details on the contents of the configuration file for the user-extended trace based performance analysis, see 7.5.3 Settings for the methods to be traced by the user-extended trace based performance analysis.
-
This is the I/O process to output the PRF trace that is output in the buffer to a file, after a certain amount of PRF trace is collected. At least one PRF daemon is invoked on each host that acquires the PRF trace. We recommend that you allocate one PRF daemon to one host.
The following figure shows the relationship between the configuration file for the user-extended trace based performance analysis and the PRF daemon.
|