Materialize put a streaming engine in a database so teams can build real-time data products without the cost, complexity, and development time of stream processing. To deliver on Materialize’s speed and correctness guarantees, we needed a strongly consistent and scalable metadata store. CockroachDB fit the bill. First, we'll walk through how CockroachDB allows Materialize to deliver on these correctness guarantees, then we'll discuss how we've leveraged CockroachDB's scalability to keep Materialize fast as we've grown our user base. We'll dive into some technical details like how using CockroachDB immediately sidestepped performance bottlenecks in vanilla PostgreSQL, and how we further tuned our data layout and access patterns for CockroachDB's architecture. Overall, CockroachDB solves hard distributed systems problems, so we can focus on our killer feature of incremental computation.