Which statement about high watermark is true?

• A. It marks the offset of the oldest message
• B. It is the offset of the last message copied to all replicas
• C. It indicates the current end offset of the log
• D. It sets the retention window for messages

Answer: (B)

High watermark represents the last offset that has been replicated to all in-sync replicas, and it defines the safe read boundary for consumers. This means consumers can rely on data up to that offset being durable across the cluster; anything beyond may still be in flight or at risk if a replica lags or fails. The high watermark advances only when every in-sync replica has acknowledged up to that offset, ensuring the data up to that point won’t be lost due to a replica failure. It is not the offset of the oldest message, nor the current end of the log, and it does not set the retention window.