Succeeding at Reactive Architecture
summary
Reactive architecture argues that asynchronous message passing and buffered brokers (Kafka, SQS, RabbitMQ) are the mechanism that unlock outbox/inbox transactional messaging, durable retries and single-threaded actors and partitioning for scale. Workflows become activities persisted via messages, enabling event-carried state transfer, eventual consistency and bulkhead resilience. Practical trade-offs covered: avoid synchronous gRPC calls, use circuit breakers and load-shedding, prefer polling for backpressure, accept at-least-once delivery with deduplication, and adopt crash-only restart semantics to recover in-flight work.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 2:05Reactive Manifesto
- 4:10Reactive programming models
- 10:31Data flow fundamentals
- 16:13Actor model benefits
- 19:45Flow-based programming
- 28:01Reactive system design
- 33:47Reactive service node
- 43:06Resilience patterns
- 46:56Crash-only recovery
- 49:23Back pressure handling
- 53:58Reactive principles recap
- 55:43Q&A
related talks
55mRethinking Reactive Architectures
34mReactive Architecture
54mReacting to an Event-Driven World
42mReactive Design Patterns
39mReacting to an Event-Driven World
42mReacting to the Future of Application Architecture
01h 00mReacting to an Event-Driven World
30mReacting to the future of application architecture