Prometheus Operators in Kubernetes - Managing and Scaling Metric Collection

Explore the challenges and solutions for managing and scaling Prometheus collection in this 28-minute conference talk by Bartłomiej Płotka and Mahmoud Amin from Google. Discover two open-source Prometheus operators for Kubernetes: Prometheus Operator and GMP Operator. Learn about their benefits, drawbacks, and how to set them up to forward metrics efficiently to remote backends like Thanos, Cortex, or cloud providers. Gain insights into different deployment models, including static Prometheus instances, scaling collections using hashmode, and dynamic scaling per node. By the end of the talk, acquire a comprehensive understanding of the trade-offs to make informed decisions on which operator to use. The speakers also discuss the potential for upstreaming functionality to a single operator in the future.