Virtual Threads - No Pain, No Gain

Explore the complexities and trade-offs of Project Loom's virtual threads in this 50-minute conference talk from GeeCON 2024. Understand how virtual threads deliver on their promise of fast, lightweight user-space threads that are cheap to block, while uncovering the hidden costs that underlying libraries must handle. Learn about the virtual thread execution model and see how it compares to the event loop/reactive model used in Quarkus. Dive deep into real-world problems that can still occur with virtual threads, including pinning, monopolization, and large thread-local objects that libraries need to account for. Discover how to verify that your code avoids these pitfalls and gain insight into issues that remain invisible from a user perspective but often manifest only under high loads, helping you make informed decisions about when and how to implement virtual threads in your applications.