Case Study: Booking.com handles millions of metrics per second with Grafana

How Booking.com Handles Millions of Metrics Per Second With Graphite

Booking.com, the Amsterdam-based travel e-commerce company, needed a way to handle massive-scale metrics for capacity planning, troubleshooting, and business visibility. As its Graphite environment grew to hundreds of servers and hundreds of terabytes of data, the team ran into scaling, reliability, and performance issues with the existing stack, including a single point of failure in carbon-relay and slow, inconsistent queries.

To address this, Grafana was installed as a more user-friendly visualization layer while Booking.com rebuilt key parts of the Graphite stack, including a custom relay, Go Carbon, and Carbon API. Grafana helped the team create more complex dashboards, and the broader overhaul improved performance dramatically: average front-end response time dropped from about 15 seconds to less than one second, query throughput rose from roughly 80 requests per second to almost 3,000 per backend, and the relay could ingest over a million data points per second using only two CPU cores.

Booking.com

Vladimir Smirnov

System Administrator