How to switch Cortex from chunks to blocks storage (and why you wonât look back)
Blog post from Grafana Labs
Grafana Labs transitioned from contributing to Cortex to launching Grafana Mimir, a scalable long-term storage solution for Prometheus, and has been focusing on enhancing Cortex's blocks engine to make it production-ready. The switch from chunks to blocks storage aims to reduce operational complexity while supporting large-scale Prometheus deployments. The migration process involves reconfiguring production clusters to store new data as blocks and converting existing data using a suite of tools called blocksconvert, which includes Scanner, Builder, and Scheduler. This conversion process results in a significant reduction in storage space by deduplicating data and optimizing label indexing. Grafana Labs has successfully migrated several production clusters with no downtime and without hindering query performance, demonstrating the efficiency and scalability of the blocks storage engine.