KIP-1352 — Ranges and Range Aggregations in the Kafka Streams DSL

Discussion Streams

Introduces a Range abstraction and range aggregations to the Kafka Streams DSL, letting an aggregation run over a set of records defined relative to an anchor record (e.g. the N events before and after it) and emit an immutable point-in-time result instead of a continuously updating KTable. Streams windowing is purely time-bucketed and incremental, so event-centric patterns equivalent to SQL's OVER clause with ROWS/RANGE BETWEEN — moving averages, ML feature vectors, threshold monitoring — have no native expression and drag in cascading updates on late data.

Details