Scaling Prometheus - How We Got Some Thanos Into Cortex

Explore a conference talk on scaling Prometheus using Cortex and Thanos. Dive into the challenges of long-term storage for Prometheus and learn about the new experimental Cortex blocks storage. Discover how this innovative approach, based on Thanos and Prometheus TSDB, aims to reduce operational costs while maintaining scalability and performance. Gain insights into the trade-offs between standard chunks storage and the new blocks storage, and benefit from lessons learned running Cortex at scale. Understand the Cortex microservices architecture, single binary mode, and the intricacies of the write and read paths. Examine performance considerations, scalability issues, and future developments in this comprehensive overview of Cortex, a CNCF sandbox project designed for multi-tenancy and high-availability in Prometheus ecosystems.

Intro
Horizontally Scalable Prometheus
Cortex Microservices Architecture
Cortex Single Binary Mode
The rise of the blocks storage
Cortex Blocks Storage Architecture
The write path: scalability issues
The write path: performances
The read path: query-frontend
The read path: querier and store-gateway
The read path: performances
The future