Hi Team, I have a scenario based question: We are using StarRocks Routine Load with Kafka. Suppose: A database backup is taken at 3 PM • The system crashes at 4 PM • Routine Load was running normally between 3 PM and 4 PM If we restore the database from the 3 PM backup: 1. How can we determine exactly which Kafka offsets were committed at the time of the 3 PM backup? 2. Are Kafka offsets included as part of the BACKUP metadata? 3. After restore, will the Routine Load job resume from the exact offsets captured at 3 PM? 4. Is there any scenario where offsets might diverge from the actual table data state? We want to ensure: • No data loss • No duplicate ingestion • Consistency between restored table data and Kafka consumption state What is the recommended recovery strategy to guarantee this?