As per the comments in the above conversation, one alternative it to write a K-Stream application. Are there any other alternatives possible?
With K-Stream, I am assuming I will have to read data from each and every topic to create an aggregated payload. This will have impact on performance. We have huge number of tables and I mean really large. I want to ensure transaction updates are done in order of the transaction. Reading on all the topics and then aggregating would come at a cost of latency and throughput. Has anyone done this? What has been the observation?
Kindly guide me.
Thank you