Debugging ETW Event Drops in High-Throughput Applications

We were losing millions of diagnostic events in production, but I couldn't touch the active monitoring agent. How do you fix a system you can't pause? In this post, I share a technique for running parallel ETW sessions to "spy" on your application's telemetry and diagnose high-frequency event drops without breaking the existing pipeline.

(1189 words)