Advanced RabbitMQ Strategies: A case study
summary
CRM platform runs >100 Kubernetes services talking via RabbitMQ (Amazon MQ) and faces unpredictable spikes, noisy neighbors, and scaling limits. Built a multicluster setup using multiple Amazon MQ brokers behind an NLB, automated via Terraform + a config registry and deployment toolkit to generate per-broker consumer deployments. Added a Java QoS service that reads headers, logs stats in Redis, detects top N offending accounts, creates dynamic offender queues, and throttles by token batches. Result: alerts/incidents drop ~70% and SLOs hold; works at day-scale spikes.
AI-written summary · about
chapters (6)
- 0:00Environment and challenges
- 2:39Problem identification
- 4:51Multi-cluster architecture
- 11:54Implementing a QoS service
- 19:49Results and conclusion
- 21:23Q&A session
related talks
40mScaling RabbitMQ
22mLessons Learned from Scaling RabbitMQ in a High-Growth (...)
32mRabbitMQ in Real Life - Tips and Tricks
32mBuilding distributed system for retail with RabbitMQ
40mBuilding Scalable Resilient IoT Messaging
01h 05mScaling Applications with RabbitMQ
35mMigrating from Kafka to RabbitMQ at SimpleBet: Why and How
30mA love letter to messaging: Celebrating RabbitMQ's journey