Case Study: Grafana Labs achieves scalable, cost-effective multi-AZ metrics architecture with WarpStream

How WarpStream Powers Grafana Labs’ Redesigned Architecture

Grafana Labs, the company behind the Grafana Mimir time series database and Grafana Cloud Metrics service, faced challenges with their existing architecture's coupled read and write paths and the prohibitively high cost of multi-availability zone deployments. Their search for a solution led them to WarpStream, as they needed a platform capable of handling tens of gigabytes per second of data throughput without incurring inter-AZ data transfer costs.

WarpStream provided a Kafka-compatible, diskless data streaming platform built directly on object storage. This solution successfully decoupled Grafana's architecture, eliminating inter-AZ networking costs and operational overhead. After validating WarpStream could meet their scale by load testing it at 1.5 GiB/s writes and 7.5 GiB/s aggregate consumption, Grafana Labs began rolling out the new, more resilient architecture for Grafana Cloud Metrics.

Grafana Labs