Case Study: Pipedrive overcomes Prometheus scalability limits with Grafana Mimir

How Grafana Mimir helped Pipedrive overcome Prometheus scalability limits

Pipedrive, the CRM platform built for salespeople, relied on Prometheus and Grafana for observability as its infrastructure grew across multiple data centers and cloud regions. But its federated Prometheus setup began hitting scalability limits, with the main instance running out of memory, taking up to 15 minutes to restart, and struggling at around 8 million active series and 20 million chunks.

To solve this, Grafana helped Pipedrive implement Grafana Mimir in a gradual, transparent migration, using remote write from Prometheus federators and tuning Mimir for their workloads. The result was a more scalable metrics platform handling 12–15 million active series, 400K samples per second, and about 600 queries per second, while reducing the risk of full-stack outages and false alerts; Grafana Mimir also gave Pipedrive better cardinality analysis and per-tenant limits.

Pipedrive

Karl-Martin Karlson

Observability Platform Team Lead