event-triggered-sagas

Kafka events trigger sagas by using an event prefix on a specific channel, allowing reactive processing to execute automatically when matching events arrive.

Kafka events can be selectively filtered using CEL expressions, ensuring that sagas only fire when specific event conditions are met.

Infinite event chains are prevented by implementing chain termination logic within the saga, stopping loops when subsequent events are emitted across services.

Idempotency is managed across saga executions by tracking state updates and ensuring reactive processing handles duplicate Kafka events without repeating side effects.

Starlark scripts are used to implement reactive saga logic after defining an event trigger and an optional filter for Kafka channels.

Debugging event-triggered sagas relies on built-in references to ADRs, manifests, and runbooks, providing auditable and traceable workflows for observing reactive behavior.