Jean-Noël Moyne
Jean-Noël Moyne
Yes, the new feature introduced in 2.8 is called deterministic subject token partitioning as an extension to the existing subject mapping functionality (which can be applied at the account level...
Thanks. I also believe that this was one of the few things some people coming from Kafka to NATS have been wanting that we didn't have until now. The feature...
Yeah I have a large doc PR in the works
FYI docs have been updated now. I think it may be ok to close this issue now?
@timsolov Of course, here is the link: https://docs.nats.io/nats-concepts/subject_mapping#deterministic-subject-token-partitioning
@rbkrabbe Partitioning does indeed allow you to scale processing while still getting in-order delivery per key, you just need to use more partition to scale further. You can have guaranteed...
Note on the terminology in use when talking about JS: 'streams' are what record messages, and you create 'consumers' on streams, those consumers are a bit like a DataBase 'view'...
You control (for example using your container orchestration system of choice) how many instances of your worker you want to have per partition. So you can have 3 partitions and...
There is no concept of "consumer groups" in JetStream where that functionality happens at the administrative/ops level and can be easily implemented using container orchestration. And I would argue that...
Just my 2 cents that I agree with: if you create a durable (either explicitly, or automatically) then it should remain after you close Drain(), Unsubscribe(), Close(), etc... a 'durable'...