Replay
The reprocessing of historical Kafka events by resetting a consumer group's offset to an earlier position. Used for re-indexing after schema changes, re-running consolidation with updated logic, and debugging pipeline failures by reproducing the exact event sequence that caused them.
Replay is one of Kafka's most operationally valuable properties. Because the topic log is durable and retains events for a configurable period, any consumer group can be repositioned to any past offset and re-process events from that point forward. In the Cognitive Substrate, replay is used in several scenarios: when a new version of the enrichment worker needs to re-index existing experience events with improved embeddings; when a consolidation bug requires re-running the consolidation pipeline over historical events; and when debugging a subtle failure by feeding the exact event sequence that preceded it back through the processing pipeline in a controlled environment. The audit stream uses a separate consumer group with offsets managed independently, ensuring that replay for debugging never contaminates the compliance record.