Drizzle is a hybrid streaming system that unifies the record-at-a-time streaming and micro-batch models. Drizzle combines the benefits from both models to achieve low latency during normal operation and during adaptation. Our key observation is that while streaming workloads require millisecond-level execution, workload and cluster properties change less frequently. Based on this insight, Drizzle decouples execution granularity from coordination granularity. Our experiments on a 128 node EC2 cluster show that Drizzle can achieve end-to-end record processing latencies of less than 100ms and can get up to 3.5x lower latency than Spark.