Smarter Golden Signals!

Platform Engineers and SREs often face alert fatigue from the sheer volume of metrics generated by Kubernetes clusters. This talk explores how Intuit tackled this challenge by implementing smarter Golden Signals using numalogic, an open-source AIOps anomaly detection engine. The presentation details how to leverage numalogic on Prometheus metrics to derive baseline behaviors and detect anomalies without requiring prior AI/ML experience. It covers the real-time collection, processing, and analysis of in-cluster data, and how numalogic computes anomaly scores to bubble up a single anomaly score for the cluster. A live demo of the AIOps-based Prometheus metrics pipeline is included.