Stream Processing has received a lot of attention first with Storm and later with Flink and Kafka. Beyond the hype, who actually use it in production and who has found it useful? What kind of use cases does it really work for?