Originally published here by Payal Chakravarty, Apr 27, 2015.
Did you know IBM Performance Management is the only solution in the market to support the full IBM Middleware stack including Websphere Application Server (WAS), Message Queue, IBM Integration Bus (IIB), DataPower, Business Process Manager (BPM), and many more. For all the large distributed applications running on IBM Middleware , you can monitor performance and availability to identify problems before they affect users, visualize bottlenecks, and find the root cause down to the line of code.
You can isolate bottlenecks across the middleware stack with automated problem detection and corrective actions for each component. You can visualize relationships and transaction flows between component
IBM APM helps Improve the availability and fine tune the performance of each application component
â€¢ WAS – Code-level diagnostics, monitor response time, thread pool connections, and more
â€¢ MQ – Queue depth monitoring for queue managers. Ensure channels are running, monitor their status and transmission queue performance
â€¢ IIB – IBM Integration Bus execution groups, visibility of message flows accounting and statistics
â€¢ DataPower – Resource utilization for appliances and Web Services metrics across domains. Monitor DataPower services traffic for response time, message counts, and more
For faster problem diagnosis across the middleware stack IBM APM offers:
â€¢ â—¦ Code level diagnostics in production on Websphere Application Servers
â—¦ Eliminate restart of the JVM when addressing hung threads
â—¦ Integrated log analytics with built-in WAS expertise
Try a few scenarios on our Live Demo: To try it for yourself, go to our Live Demo and click on “Online Catalog” application
1. Detect Problems with Resource Monitoring of IBM Middleware Components
You will notice from the “Current Component Status” of this application that it is running on several IBM middleware components: IBM HTTP Server, WebSphere Application Server, DB2, MQ, IIB, Datapower.
On the left hand navigation, select Groups -> Components. At one glance you can see the high level health/status of all the middleware components that the application runs on and detect the problem areas. For example you can see there are some WAS log errors, Message Queue has some issues with channels and transmission queue messages. You can click on any of these to drill in to diagnose the component in trouble.
2. Isolate Problems in the middleware Java stack with Transaction Tracking
To walk through a problem isolation scenario select the “View Transaction Details” guided demo on the Live Demo.
3. Diagnose Problems to the line of J2EE code running on WAS
Once you have isolated which transaction instance is causing a problem (in the previous step), we will move to the diagnsotic phase.
On selecting a specific transaction instance, the summary information for the transaction instance and a topology of the instance are shown. The topology shows the transaction instance organized into network hops and subtransaction nodes. Notice that the icons show failed status for the nodes and the time taken for each hop is provided. Select the WebSphere Application Server node to display further information about it.
Click the Diagnose link below the Node Properties table to see all the Request Instances. Click “View Request Sequence” to drill down and see the method and stack trace of the problem. IBM APM will keep track of the response time trends and start capturing stack traces as soon as the response time exceeds the dynamic baseline.
On expanding the stack trace you can see the exact JDBC query that was taking long to execute and hence caused a failure. You can eliminate restart of the JVM when addressing hung threads.
For more information check out the “Performance Management for IBM Middleware” page on Service Engage. https://www.ibmserviceengage.com/application-monitoring/articles/apm-for-IBM-Middleware