AIOps: Anomaly Detection with Prometheus and Istio
summary
Shows a working pipeline that scrapes Prometheus, stores long term with Thanos long-term storage or JSON blobs on S3, and analyzes data with Apache Spark and the Open Data Hub. Forecasting uses Facebook Prophet to produce upper and lower bounds, predictions are exposed on a /metrics endpoint so Prometheus scrapes them, and alerts fire when an accumulator threshold marks a deviation. Demo runs on OpenShift, handles about 360,000 metrics per hour, and retrains hourly.
AI-written summary · about
chapters (4)
related talks
37mAIOps: Anomaly Detection with Prometheus and Istio
40mAIOps: Anomaly detection with Prometheus
27mObservability with Prometheus and beyond
31mHandling Billions of Metrics with Prometheus and Thanos
42mAIOps for Distributed Environments - Deep Dive
35mAnomaly Detection for Cloud Native Storage
24mObservability in the MLOps Lifecycle with Prometheus
29mParanoid Observability –Success Stories and Misfortunes observing an observability stack