Session abstract
What you’ll learn
Serverless architectures are great... until they aren't. When your application demands long-running, multi-step processing and built-in resilience, a simple serverless function won't cut it. You'll need durable executions and stateful processing, but what is it, and how can it help? In this talk, we'll look at fault-tolerant, durable execution patterns. We'll cover checkpointing, replay, and how step functions work. To make this concrete, I'll walk through an order system that authorizes a payment, waits for manual approval, then captures or cancels the charge. We'll dig into the code and see what improvements this has over traditional serverless functions. By the end, attendees will have a better idea on how durable executions work, and why they should use them.