Sophie, Thank you for the KIP! Choosing the number of partitions in a Streams app is a tricky task because of how difficult it is to re-partition; I'm glad you're working on an improvement. I've got two questions:
First, `StaticStreamsPartitioner` is an interface that we (Streams users) must implement, I'm trying to understand how it would work. For example, let's say there's some point in time 'T' before which we have 1 partition. Then we decide to increase the partition count to 2 at time T. From my understanding, all keys that had passed through the Streams app before time T must end up on partition 1 if they appear again in the input topics; but any new keys are allowed to be sent to partition 2. Is that correct? And (pardon the naive question) how is this achieved without keeping track of all keys that have been seen at any point? Secondly, will this feature work with applications that use interactive queries? Thank you very much, Colt McNealy *Founder, LittleHorse.io* On Tue, Oct 18, 2022 at 9:34 PM Sophie Blee-Goldman <sop...@confluent.io.invalid> wrote: > Hey all, > > I'd like to propose a new autoscaling feature for Kafka Streams > applications which can follow the constraint of static partitioning. For > further details please refer to the KIP document: > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-878%3A+Autoscaling+for+Statically+Partitioned+Streams > > This feature will be targeted for 3.4 but may not be fully implemented > until the following release, 3.5. > > Please give this a read and let me know what you think! > > Cheers, > Sophie >
