Using the service logger to identify performance issues

Edit online

To get better performance, as measured in response time and reliability, review statistics and modify properties and parameters to improve results.

About this task

To identify the source of any performance problems:

Procedure

Enable the service log by enabling the corresponding java loggers. Edit either of the following files to change the trace settings:
- Websphere-Liberty-Profile-Folder/servers/defaultServer/server.xml
- /SI/properties/meglogging.properties
To enable tracing for Global Mailbox, append one of the following strings to traceSpecification:
- Entry tracing: com.ibm.service.entry*=fine
  - Monitor main application entry points
  - Negligible performance impact
- Light Application Layer Tracing: com.ibm.service.*=fine
  - Monitor the main application layers
  - Doesn't include result size measurement
  - Light performance impact
- Complete Application Layer Tracing: com.ibm.service.*=finest
  - Monitor the main application layers, including result size and entry logs
  - Average performance impact
- Complete Tracing: com.ibm.service.*=finest:com.ibm.mailbox.*=finest
  - Monitor every entry and exit calls during application execution
  - Significant performance impact
For example:
```
com.ibm.service.level=fine
```
Tip: This can be done while the server is running.
As the application runs, the logs are gathered in the file : (Websphere-Liberty-Profile-Folder)/servers/defaultServer/logs/trace.log.
The following text is an example of the log file:
```
[2014-08-28T08:38:05.600-0700] 00000001 LogMetricGath 3 com.ibm.commerce.cache.LogMetricGatherer 
```
```
logMetricToLogger <exit operation="Operation : com.ibm.mailbox.api.repository.IMailboxService.getMailbox(UUID
```
```
mailboxId)" parameters="00000000-0000-0000-0000-000000000000," id="338020115243073538" parentId="0" 
```
```
startTime="4518568325358" stopTime="4518584207505" duration="15882147" resultSize="517" cacheHit="false"
```
```
cacheEnabled="false" />
```
The definitions for the fields in the log are the following descriptions:

operation

Name of the operation that was executed

parameters

Parameter values that were sent during the operation execution. Note that some parameter values might be hidden if they contain sensitive information.

id

Unique operation execution identifier. Used to correlate operations together.

parentID

Unique identifier of the parent operation execution. Used to correlate the hierarchy of operations.

startTime

Start time value in nanoseconds in a value that is relative to the stopTime value.

stopTime

Stop time value in nanoseconds in a value that is relative to the startTime value.

duration

Duration of the operation execution in nanoseconds.

cacheEnabled

Boolean flag that indicates if result caching was enabled on this operation.

cacheHit

Boolean flag that indicates if a cache hit was performed.
Review the service logs for important entry and exit events and important messages.
You can learn the following types of information from the log:
- Average execution of each operation in the important application layers
- Average response size
- A list of all the request received by the server sorted by response time (slowest ones show up first)
- The hierarchy of calls
- (Advanced) information about cache potential of each request
Use this information to monitor the performance of your system and resolve problems or improve response.