7.7.1 Overview of the Operation for Analyzing the Processing Performance
You can analyze the processing performance of Application Server based on the trace information output from the function layer such as the EJB container and Web container, in the series of processes from the client to EIS such as databases, until the processing result is returned to the client.
Note that the trace information of each function layer is output to the PRF trace file in the binary format. When analyzing the processing performance of Application Server, use the trace based performance analysis file where PRF trace file is converted in to text format (CSV format).
The description of the operation of analyzing the processing performance of Application Server is as follows:
- Reference note
-
Maintenance level trace information
The PRF trace collection level includes the maintenance level in addition to the standard level and detailed level. The maintenance level is a level for acquiring the maintenance information in the case of an error. Do not specify the maintenance level in normal cases.
For how to set up the maintenance level in PRF trace collection levels, see cprflevel (display or change the PRF trace collection level) in the uCosminexus Application Server Command Reference Guide.
- Organization of this subsection
(1) Collecting a trace based performance analysis file
Collect the trace based performance analysis files used for analyzing the processing performance of Application Server, which is output using the management command mngsvrutil. For details on how to collect the trace based performance analysis file and on the output destinations and output information of the trace based performance analysis file, see 7.3 Collecting the trace based performance analysis file by using Management Server.
(2) Analyzing the processing performance of Application Server using a trace based performance analysis file
To analyze the performance using a trace based performance analysis file, display the file in an application program that can edit CSV format files, and use the filtering or sorting function as required.
For example, when the information is output in the CSV format, you can filter the output items as 'event-wise', or 'process-wise', in the application programs that can edit files in the CSV format, list the locations to be focused, and analyze the information according to your requirement.
How to use the trace based performance analysis file is described below by giving an example.
-
For details on analyzing the response time of the Web server, see 7.7.2 Analyzing the Response Time of a Web Server.
-
For details on investigating the processing status of a request in Application Server, see 7.7.3 Investigating the Processing Status of a Request in an Application Server.
-
For details on investigating the life cycle of the session, see 7.7.4 Investigating the Life Cycle of a Session.
-
For details on identifying the transaction in which the timeout occurs, see 7.7.5 Identifying the Transaction in Which a Timeout Occurred.
-
For details on identifying of the request for which time-out occurs, see 7.7.6 Identifying the Request for Which Timeout Occurred.
-
For details on the log investigation where root application information is used, see 7.7.7 Investigating the Log Using the Root Application Information.
-
For details on identifying the connection where the trouble occurred, see 7.7.8 Identifying the Connection in Which an Error Occurred.
-
For details on the investigation about the location of the problem associated to Trace based performance analysis and thread dump, see 7.7.9 Investigation about the location of the problem associated to the trace based performance analysis file and thread dump.
Note that the collection of trace based performance analysis files in such an environment as the following Web client configuration environment is a prerequisite.
|