Example: Monitoring service component performance

Why and when to perform this task

You will use the business rules sample application for this scenario, where you will monitor all three of the performance statistics: successes, failures, and response times. You should have the web page containing this application already open; keep it open, since you will be running the sample several times after you begin monitoring. Ensure that you have already run the sample at least once, so that it will appear in the list of functions that you can select to monitor.

Steps for this task

  1. Open the administrative console.
  2. Click Servers > Application Servers in the console navigation tree.
  3. Click server_name.
  4. Click the Runtime tab.
  5. Click Performance Monitoring Infrastructure under Performance.
  6. Select Custom.
  7. Expand WBIStats.RootGroup > BR > brsample_module.DiscountRuleGroup > Operation, and select _calculateDiscount.
  8. Check the boxes next to BadRequests, GoodRequests, and ResponseTime, and press Enable.
  9. Click Monitoring and Tuning > Performance Viewer > Current Activity in the administrative console navigation tree.
  10. Check the box next to server_name and press Start Monitoring.
  11. Click server_name.
  12. Expand WBIStats.RootGroup > BR > brsample_module.DiscountRuleGroup > Operation, and check the box next to _calculateDiscount.

Result

You should now see a blank graph, and underneath that the names and values for the three statistics. Check the boxes next to the statistic names, if they are not already checked. The PMI is now ready to publish performance data for the selected event, and the TPV is ready to present the results.

Run the business rules sample application several times, and then watch the performance viewer as it periodically refreshes. Notice that there are now lines on the graph, representing the cumulative number of successful requests and the average response time for each successful request. You can also see the values next to the name for each statistic below the graph. The line for the number of successes should continue to rise as you perform additional invocations of the sample, while the response time line should level off after a few refreshes.

Once you have completed this task, you should understand how WebSphere® Process Server implements performance monitoring of service components. You will know how to select service components for monitoring, and how the performance statistics are calculated. You will also be able to start the performance monitors, and view the performance measurements for your applications as they are being used.

What to do next

Performance monitoring can tax system resources, so after you have completed this task you should stop the monitors. To do this, simply click on the Tivoli® Performance Viewer link, select both the node and the server, and press Stop Monitoring.

Terms of use |

Last updated: Tue Feb 21 17:02:05 2006

(c) Copyright IBM Corporation 2005.
This information center is powered by Eclipse technology (http://www.eclipse.org)