Explain the producer request lifecycle with acks=all in detail.

1) Producer serializes record and determines target partition (key hash or round-robin). 2) Record is appended to RecordAccumulator buffer. 3) Sender thread batches records and sends ProduceRequest to partition leader. 4) Leader writes records to its local log (LEO advances). 5) Leader waits for all ISR followers to send FetchRequests and acknowledge the new offset. 6) HWM advances to the new offset once all ISR acks are received. 7) Leader sends ProduceResponse with offset to producer. 8) On ACK timeout or retryable error, producer retries per retry config.

Expected key concepts: RecordAccumulator, Sender thread, ProduceRequest, leader, local log, LEO, ISR followers, FetchRequests, HWM, ProduceResponse, retry