Configuring the IIS Request-Based Tracing Session
Applies To: Windows Server 2003, Windows Server 2003 with SP1
Setting up a request-based tracing session in IIS requires that you first decide what you want to trace. Your decision on what to trace will be determined by the nature of your Web applications. For example, if your Web applications use Active Server Pages (ASP), then you would list this provider in the provider file. If your Web applications do not use ASP, then you would not want to list this provider in the provider file. After you start tracing, the components defined in the provider file report trace data to the ETW infrastructure.
Your decision on what to trace will also be determined by the nature of the problem you are troubleshooting. If you know which component is causing the problem you can create your provider file to target the problematic component. If you are having troubles with a specific application, you can trace the specific URL.
If you are not certain which URL or component(s) are causing the problem, use IIS currently executing requests to determine which URL is being requested when a worker process becomes unresponsive. Knowing which URL is causing the problem will also help you determine which other providers to define in the provider file. For example, if you know the problem exists in a URL that utilizes an ISAPI filter, then you would want to define that provider in the provider file. By explicitly defining which components to trace, you reduce the amount of data in the trace log, which makes troubleshooting easier.
This section discusses the following concepts related to setting up an IIS request-based trace session: