It's worth noting that most managed Kafka solutions offer a service which will transform a stream into a database table so you can query it later. We've also bu

It's worth noting that most managed Kafka solutions offer a service which will transform a stream into a database table so you can query it later. We've also built our own tooling to do it. We read from the Kafka stream directly when we're building a pipeline between microservices; for almost everything else we query the database. It works pretty well and we haven't found ourselves wishing we could write pseudo-queries on streams yet.