Thread deadlocks are a notorious problem in ASP.NET applications, often leading to complete service unavailability. This post provides a systematic methodology for diagnosing such issues using WinDbg, a powerful Windows debugger. The approach begins with high-level thread pool metrics using the !tp command to identify signs of thread starvation—such as all threads being busy with zero idle threads and a growing queue. Next, it guides the developer to dump all thread call stacks with ~*kb to locate threads stuck in a waiting state. By analyzing synchronization objects and lock hierarchies, the root cause can be identified. The post also covers common patterns like lock ordering violations and thread pool exhaustion due to blocking asynchronous calls. For backend developers and system administrators, mastering this diagnostic process is essential for maintaining high availability in production .NET environments. The methodology is reusable across different versions of ASP.NET and can be adapted for other .NET application types.
This post presents a standardized approach to diagnosing thread deadlocks in ASP.NET applications using WinDbg. It outlines a step-by-step methodology starting with high-level thread pool metrics, identifying blocked threads, and analyzing call stacks to pinpoint the root cause. This is a valuable resource for backend developers dealing with production issues, as thread deadlocks are a common and hard-to-debug problem in .NET applications.