HI everyone, I'm trying to understand the relationship between fetch.max.bytes and replica.fetch.response.max.bytes in Apache Kafka. According to the KIP-541 documentation <https://cwiki.apache.org/confluence/display/KAFKA/KIP-541%3A+Create+a+fetch.max.bytes+configuration+for+the+broker> , fetch.max.bytes is the maximum fetch request size from replicas, and it seems to act as a cap.
I'm concerned that if replica.fetch.response.max.bytes is set larger than fetch.max.bytes, it could lead to issues, such as high CPU usage on the replicated broker when the fetch.max.bytes limit is hit. This behavior is also suggested by the code here <https://github.com/apache/kafka/blob/3.8.1/core/src/main/scala/kafka/server/KafkaApis.scala#L1033> . Could someone with more experience confirm if this understanding is correct and whether it would be beneficial to add a note about this relationship in the documentation? Thanks! Regards
