Hello, I have an application that reads from two Kafka sources, joins them, and produces to a Kafka sink. The application is experiencing long end to end checkpoint durations for the Kafka source operators. I'm hoping I could get some direction in how to debug this further.
Here is a UI screenshot of a checkpoint instance: [image: checkpoint.png] My goal is to bring the total checkpoint duration to sub-minute. Here are some observations I made: - Each source operator task has an E2E checkpoint duration of 1m 7s - Each source operator task has sub 100ms sync, async, aligned buffered, and start delay - Each join operator task has a start delay of 1m 7s - There is no backpressure in any operator These observations are leading me to believe that the source operator is taking a long amount of time to checkpoint. I find this a bit strange as the fushioned operator is fairly light. It deserializes the event, assigns a watermark, and might perform two filters. In addition, it's odd that both source operators have tasks with all the same E2E checkpoint duration. Is there some sort of locking that's occurring on the source operators that can explain these long E2E durations? Best, Hubert
