High availability dalam RabbitMQ melibatkan clustering (berbilang node) dan queue yang direplikasi (mirrored atau quorum queue) supaya sistem bertahan daripada kegagalan node tanpa kehilangan mesej. Memahami HA adalah penting untuk RabbitMQ yang boleh dipercayai dan produksi.
Clustering
RabbitMQ CLUSTER → multiple RabbitMQ nodes working together as one logical broker:
→ nodes share metadata (queues, exchanges, bindings, users)
→ distributes connections and load across nodes
⚠️ by default, a queue lives on ONE node → if that node fails, the queue is unavailable
(clustering alone doesn't make queues highly available → need replication)
