Hello, I was trying to understand the effect of consumer liveliness and use of resources at the same time, based on this article (which is great, btw):
https://www.confluent.io/blog/tutorial-getting-started-with-the-new-apache-kafka-0-9-consumer-client/ For my use case, I have only one consumer per group which is used in the following way: 1) Instantiated with all properties. 2) poll messages 3) Process all messages and handles error 4) Commits offsets in sync.(no auto commit) 5) Closes all resources. The above means that I will have a rebalance every time a consumer joins as part of this consumer group. Because there is only 1 topic and 1 partition, its not too troublesome at the moment. I would like to exploit multiple partitions for scaling purposes. But I understand that having multiple consumers doing the above steps means frequent rebalancing. But if Isimply process everything in a while loop (until the server goes down), the resources are never closed until a clean shutdown. I was curious to know whether there is a practical answer to this. Additionally, if anyone has implemented this across multiple containers/servers in cloud, it would be good to know. Also, apologies if any incorrect understading/assumption has been made. Regards,
