Migrating from Kafka to RabbitMQ at SimpleBet: Why and How
summary
SimpleBet replaces a Kafka-per-customer design with a three‑node RabbitMQ cluster exposing vhost per customer sandboxes, AMQP auth, management API provisioning and JSON messages, cutting infrastructure and customer onboarding. The system uses exchanges, shovels and headers to route and filter feeds, durable/quorum queues for persistence, and Elixir publishers (amqp library) with occasional ex_rabbit_pool. Peak volume is tiny—one‑two messages/sec—so RabbitMQ meets low‑latency, low‑cost, and easier‑to‑integrate goals and is running in production.
AI-written summary · about
chapters (8)
- 0:0004:00 - Introduction
- 4:0104:51 - The first steps at Simplebet and the problem
- 4:5205:42 - Serving Odds
- 5:4306:24 - Market Publishing Requirements
- 6:2515:22 - Kafka
- 15:2316:28 - Re-evaluating our Integration
- 16:2926:55 - RabbitMQ enters the picture
- 26:5635:00 - QnA
related talks
24mAdvanced RabbitMQ Strategies: A case study
30mA love letter to messaging: Celebrating RabbitMQ's journey
01h 00mRabbit MQ vs Kafka: When, Which, Where?
32mRabbitMQ in Real Life - Tips and Tricks
39mReal World RabbitMQ: Exploring Practical Applications
39mScaling Betfair Exchange to support 160k rps using Apache Kafka
40mScaling RabbitMQ
36mBeyond the Burrow: The Endurance of RabbitMQ