Overview
Explore the powerful world of Event Tracing for Windows (ETW) in this comprehensive 58-minute conference talk. Learn how to diagnose complex issues in production systems, including excessive garbage collection pauses, slow startup, and Windows boot process sluggishness. Discover techniques for automating ETW collection and analysis to build self-diagnosing applications that identify high CPU issues, resource leaks, and concurrency problems. Gain insights into innovative performance tools like flame graphs for visualizing call stacks and a command-line interface for dynamic, scriptable ETW tracing. Master this critical skill for application performance and diagnostics, covering topics such as CPU hot paths, garbage collection, memory traffic, concurrency patterns, and real-time analysis with TraceEvent. Understand the architecture of ETW, learn to collect and parse event data, and explore Windows Performance Recorder for monitoring production systems and boot processes.
Syllabus
Intro
Challenges with Traditional Profilers
Event Tracing for Windows
Architecture Overview
CPU Hot Paths
Garbage Collection
Memory Traffic
Concurrency Patterns
Collecting Data
Managed Stacks
EventParsers
Real-Time Analysis with TraceEvent
Monitoring a Production System
Windows Performance Recorder
What's Running During Boot?
Main Takeaways
Summary
Taught by
NDC Conferences