Which type of watermark is automatically managed by the system?

The concept of watermarks in data processing environments refers to markers that indicate the progress of data processing or the state of data consumption in a streaming or batch context. The choice of automatic watermark is correct as it refers specifically to those watermarks that are managed by the system without requiring manual intervention.

Automatic watermarks are generated by the underlying framework based on the processing of incoming data. This method helps in efficiently tracking the point up to which data has been processed, ensuring that the system can continue transactional integrity and fault tolerance without user input. The system can adjust these watermarks dynamically, making it suitable for scenarios where data arrival is unpredictable or variable.

This contrasts with other types of watermarks that require manual oversight or are transient and only relevant for temporary states of processing. By having automatic watermarks, the system enhances its robustness and maintainability by minimizing the potential for human error while ensuring consistency in data handling.