Currently, Kafka source connector is using changeFeedProcessor to pull changes. Will raise few challenges: 1. Offset tracking 2. load balancing ChangeFeed pull model will be a better suite.