Rill, an operational BI-as-code tool, uses ClickHouse as its analytical engine to process 100B+ daily events for customers like Bloomberg, AT&T, and Comcast. The architecture combines declarative data ingestion via dlt, high-performance aggregation in ClickHouse, and a metrics-first semantic layer where dashboards are generated from code rather than built manually. Developers define sources, metrics, and dashboards in SQL and YAML, iterate locally, version through Git, and deploy like software. The system also supports conversational analytics with natural-language queries backed by ClickHouse's speed, avoiding the latency issues seen with Snowflake or BigQuery. Each AI-generated insight links back to underlying queries for verifiability.
Table of contents
A fast BI tool for real-time databases #Inside Rill's ClickHouse-based architecture #How BI-as-code works in practice #Analytics for a fully metered world #Sort: