Case Study: Character.AI cuts observability costs by 50% with ClickHouse ClickStack

A ClickHouse Case Study

Preview of the Character.AI Case Study

Scaling observability at Character.AI thousands of GPUs, 10x logs, and 50% lower cost with ClickStack

Character.AI, an AI platform serving millions of active users and running thousands of GPUs across multiple Kubernetes clusters and cloud providers, needed a scalable way to centralize observability. Its logs were fragmented across systems, queries were slow, and costs were hard to predict. ClickHouse and ClickStack were chosen to provide a unified logging and search experience for the company’s high-volume production environment.

With ClickHouse Cloud and ClickStack, Character.AI deployed a fast, centralized observability stack that ingests logs directly from OpenTelemetry collectors, supports powerful log search, and scales automatically. The result was a 10x increase in log data handled while cutting costs by 50%, with search times dropping from minutes to seconds and long-retention queries becoming dramatically faster.


View this case study…

Character.AI

Mustafa Yildirim

Reliability Engineer


ClickHouse

121 Case Studies