Glossary
Message Broker
A message broker is middleware that receives messages from producers, routes them based on rules, and delivers them to the appropriate consumers. It acts as the central hub in event-driven and microservice architectures, decoupling producers from consumers.
General definition
A message broker sits between the services that create messages and the services that consume them. Instead of services communicating directly, they publish to the broker, which applies routing logic (via queues, topics, or exchanges) and forwards messages to one or more consumers. This central coordination point adds reliability, flexibility, and observability to distributed systems.
- Topic-based routing so different consumers receive different subsets of events
- Message persistence and replay when consumers are offline
- Protocol bridging between producers and consumers using different formats
- Dead-letter handling, retries, and delivery tracking
Well-known brokers include RabbitMQ (which speaks AMQP), Apache Kafka, and cloud services such as AWS SNS/SQS. In contrast to a pub/sub bus, a broker typically persists messages and tracks acknowledgements explicitly.
In the Ethora ecosystem
Ethora’s server-side architecture uses broker-style event routing for push notification delivery, webhook fan-out, and AI agent job dispatch. When you run Ethora on your own infrastructure, you can connect your existing enterprise broker, giving your ops team full visibility into message flows and replay capability if a downstream system fails.
For compliance-sensitive deployments, a self-managed broker provides a durable audit trail of every event that passes through your chat and AI platform, which matters for healthcare and financial services customers who need message-level traceability.
