Observability is essential for ensuring the scalability and performance of modern platforms. This talk will demonstrate how to monitor thousands of targets using Prometheus and Thanos, with a focus on scalable remote storage and long-term retention. Attendees will explore dynamic service discovery, metric aggregation, and multi-cluster monitoring setups. Real-world examples will illustrate techniques for maintaining platform performance while managing cost-effective observability. This session is designed for platform teams managing complex, large-scale systems.