Hi, Wanted to understand how KafkaSource stores the offsets of each partition in state. Does it keep the offsets of each topic-partition in its checkpointed state or just commit back to Kafka on every checkpoint? And then how does DynamicKafkaSource work in this? Does it store the offsets for all cluster-topic-partitions in state? If so, will it restore this state to find offsets for a cluster if it has been dynamically readded? Would appreciate if someone can explain this. Thanks in advance. Chirag
