Data & Analytics · Open-source build
Real-time product metrics on ClickHouse
Events from the product arrive by the million and are queryable within seconds, on hardware a small team can afford.
What it is
Product events stream through Kafka into ClickHouse, where materialised views maintain rollups as data arrives. Dashboards query the rollups and drill into raw events when needed. Retention tiers move old partitions to cheaper storage automatically.
What the build consists of
- 01Event schema with versioning and a validating collector
- 02Kafka topics into ClickHouse with exactly-once semantics via Kafka engine tables
- 03Materialised views for funnels, retention and sessions
- 04Tiered storage: hot on SSD, warm on object storage
- 05Grafana dashboards with sub-second queries over billions of rows
What you see in a demo
A load generator producing tens of thousands of events per second while a funnel chart updates live, and a query over the full history returning before you finish reading it.