Summary
Kafka consumer groups implement backpressure via the consumer poll loop: pausing the poll loop stops consumption without dropping messages, providing durable backpressure to the producer.
Evidence
- ·backpressure implementation_notes: monitor Kafka consumer-group lag, pause() the consumer above a threshold and resume when lag drops (patterns/backpressure.yaml)
- ·kafka documents consumer lag accumulation when consumers cannot keep pace with producers (technologies/kafka.yaml bottlenecks)
Evidence grounding
Grounded, 2 supporting itemsKafka backpressure is applied downstream by pausing the consumer poll loop on lag.