This blog post discusses the new feature of on-demand state repartitioning in Apache Spark™ Structured Streaming on Databricks, which aims to optimize the management of stateful streams. It walks through the implementation details, benefits of the feature, and how it can influence production workloads.