High Available and Scalable Prometheus with Thanos in Alibaba

Explore how Alibaba Group leverages Kubernetes, Prometheus, and Thanos to support their massive e-commerce operations in this conference talk. Discover the challenges and solutions for implementing a highly available and scalable fine-grained monitoring system capable of handling 4 million TPS and 10K requests per second. Learn about scaling Prometheus for large-scale scenarios, optimizing query latency across multiple Prometheus instances using Thanos, and gain insights into effective configuration practices for target discovery, recording rules, and alerting rules. Delve into the experiences and lessons learned by Alibaba's team in developing a robust monitoring infrastructure for their cluster management system.