The dimensions of monitoring

 

Performance problems can be almost anywhere. The problem can be network and hardware related, backend system related, it can be actual product bugs, or quite often, application design issues.

Understanding the flow used to diagnose a problem helps to establish the monitoring that should be in place for your site to detect and correct performance problems. The first dimension is the "end-user view" - the black box view of your Web site. This is an external perspective of how the overall Web site is performing from an end users view and identifies how long the response time is for an end user. From this black box perspective, it is important to understand the load and response time on your site. To monitor at this level, many industry monitoring tools allow you to inject and monitor synthetic transactions, helping you identify when your Web site experiences a problem.

The second step is to understand the basic health of all the systems and network that make an end user request. This is the "external" view which typically leverages tools and utilities provided with the systems and applications running. In this stage, it is of fundamental importance to understand the health of every system involved - including Web servers, appservers, databases, backend systems, etc. If any of the systems has a problem, it may have a rippling effect and cause the "servlet is slow" problem.

This dimension corresponds to the "what resource is constrained" portion of the problem diagnosis. To monitor at this level, WebSphere provides PMI instrumentation and the Tivoli Performance Viewer as a starting point. There are also several industry tools built using PMI instrumentation that provide 24x7 monitoring capabilities.

The third dimension is the application view. This dimension actually understands the application code that is satisfying the end user request. This dimension understands that there are specific servlets that are accessing session beans, to entity CMP beans, to a specific database, etc. This dimension typically comes into play in the in-depth internal understanding of who is using the resource. Typically at this stage, some type of time trace through the application, or thread analysis under load conditions techniques are deployed to isolate areas of the application, and particular interactions with backend systems or databases that are especially slow under load. WebSphere provides the Request Metrics technology as a starting point. In a lot of cases, you start moving into using a lot of the development tools provided such as IBM WebSphere Studio Application Developer.

  Prev | Home | Next

 

WebSphere is a trademark of the IBM Corporation in the United States, other countries, or both.

 

IBM is a trademark of the IBM Corporation in the United States, other countries, or both.

 

Tivoli is a trademark of the IBM Corporation in the United States, other countries, or both.