The Simplest Way to Build Resilient Applications
summary
Concrete claim: durable execution removes glue code by persisting an execution log so retries replay deterministically and side effects aren’t repeated. Live demos use Java/Spring handlers and Restate durable execution engine (single Rust binary, S3 snapshots) with the SDK calling ctx.run deterministic actions and ctx.awaitable durable futures to implement idempotent payments, saga compensations via a terminal exception, Kafka-driven parallel claim processing per key, and a package-tracker state machine — all recover and resume correctly in the demos.
AI-written summary · about
chapters (10)
- 0:00Introduction and context
- 1:04Challenges in microservices
- 5:46Introducing durable execution
- 8:11Workflow and demo walkthrough
- 15:14Implementing sagas
- 19:24Async tasks and timers
- 23:23Kafka event processing
- 27:44Stateful entity management
- 30:26Architecture and summary
- 35:31Q&A session
related talks
48mThe simplest way to build resilient applications
45mThe simplest way to build resilient applications
35mBuilding resilient applications with Durable Execution
27mThe simplest way to build resilient applications
53mDurable Execution With Restate
38mDurable Execution: The Next Leap for Microservices Architectures
28mStateful Applications in the Age of Durable Execution
06mDurable Executions in the Face of (some) Failures