I have one flink job which consume 2 Kafka stream Application and transaction. Both has almost 1 year data. Based on which I am computing something and pushing a values. Now I need to backfill both whole stream and compute hisorical features. trigger event is application based on which i compute feature by taking value of transaction. Which I successful able to do by using CoProcessFunction function
Problem is application stream is very light weight has less data compare to transaction. now I want to hold application stream untill all the data has been backfilled to my state of transaction then i want to proceed for application, how to do it?