[ Question - Routine Load Throughput ] Hi Folks! I currently have 3 CN nodes, and when using Routine Load to ingest data from Kafka, each job creates a maximum of 3 tasks (one per CN). Because of this limitation, the overall throughput of a single Routine Load job is capped, and Kafka lag is increasing. To increase ingestion throughput without adding more CN nodes, would the recommended approach be to split the Kafka partitions into multiple Routine Load jobs? For example, dividing 48 partitions into 3 or more jobs so that each job runs 3 tasks in parallel, allowing the cluster to process more partitions concurrently. Additionally, I’m considering enabling

property.auto.commit

.From the Kafka UI, it appears that lag is accumulating, and I’m wondering if enabling auto commit is safe for Routine Load, or if it might lead to data consistency or offset-management issues. What is the best practice in this scenario? 🤔