...

For example, MySQL CDC source should report isProcessingBacklog=true at the beginning of the snapshot stage. And it should report isProcessingBacklog=false at the beginning of the snapshot changelog stage.

Example Usages

Let's re-use the example use-case specified in the motivation section. Suppose a job needs to consume data from HybridSource composed of an HDFS source and Kafka Source (or MySql CDC Source),  and produce data into Paimon Sink (or any sink with exactly-once semantics). Users want to trigger the checkpoint once every 30 minutes during the first phase so that the job will re-process at most 30 minutes worth of work after failover; then trigger the checkpoint once every 30 seconds during the second phase so that the sink will flush/commit/expose data to downstream applications once every 30 seconds (needed for data freshness)

...