Which statement about the high watermark is correct?

Which statement about the high watermark is correct?

• A. The high watermark is the offset of the last message that has been committed to the topic
• B. The high watermark is the offset of the last message that was successfully copied to all replicas
• C. The high watermark is the offset of the newest message produced to the topic
• D. The high watermark marks the end of the log

Answer: (B)

The high watermark defines the boundary of data that Kafka considers safely committed. It is the offset of the last message that has been replicated to all in-sync replicas, so those messages are durable and readable by consumers. Because some followers can lag, the high watermark sits behind the log’s end offset, marking what’s guaranteed to exist on all in-sync copies. Consumers can read up to this point, but not beyond it until those newer messages catch up. This is why the statement describing the high watermark as the offset of the last message copied to all replicas best fits the concept. The newest produced message is not the high watermark, and the high watermark is not the literal end of the log.

The high watermark defines the boundary of data that Kafka considers safely committed. It is the offset of the last message that has been replicated to all in-sync replicas, so those messages are durable and readable by consumers. Because some followers can lag, the high watermark sits behind the log’s end offset, marking what’s guaranteed to exist on all in-sync copies. Consumers can read up to this point, but not beyond it until those newer messages catch up. This is why the statement describing the high watermark as the offset of the last message copied to all replicas best fits the concept. The newest produced message is not the high watermark, and the high watermark is not the literal end of the log.