Hi All,
In one of the Flink jobs, the JobManager is checkpointing every 15 seconds, and the checkpoint file size is close to 1GB usually. I’m debugging an issue related to this job, where the checkpoint size was doubling every checkpoint and eventually crashing after reaching ~20GB.
I suspect that an input message with a future timestamp might have caused this issue. Could it be the cause? Any pointers or articles on this topic would be really helpful. Thanks.