ClickHouse

Why we included this project

ClickHouse is a column-oriented database that answers analytical queries over very large tables in near real time, and that speed is what earns it a role in ML pipelines. Teams use it for training-set preparation, feature stores, and embedding lookups, where you need aggregations and point-in-time scans without waiting minutes for a result. It also works as a general analytics warehouse for telemetry, logs, and other event data, so the same deployment can serve both the ML data layer and everyday reporting. The project is mature enough to run self-hosted at serious scale, installs with a single command, and spares you from gluing together several specialized systems when one fast analytical store will do.