Event-driven architecture (EDA)

Event-driven architecture is a design style where components communicate by emitting events, facts about something that happened, rather than by calling one another's functions or APIs directly. An order service publishes an "order placed" event; a shipping service, an email service, and an analytics pipeline each independently notice and react to it, with no direct dependency between the services.

This is a fundamentally different coupling model from request-response architectures, where a service calls another and waits for a reply, and knows exactly who it's calling. In EDA, producers don't know or care who, if anyone, is listening, and consumers can be added or removed without changing the producer. Events typically flow through a message broker or an event streaming platform such as Apache Kafka, and simple point-to-point notifications are often delivered as webhooks.

EDA matters because it lets systems evolve and scale independently, a new consumer can subscribe to existing events without any change upstream, and it fits naturally with stream processing for reacting to data in near real time. The tradeoffs are real: debugging is harder because there's no single call stack to trace, and ensuring consistency across services that each react to events at their own pace, rather than within one transaction, requires deliberate design.