AIOps: Anomaly detection with Prometheus
summary
Shows how Prometheus metrics are exported to long-term storage (Thanos or Influx or raw JSON blobs on object storage) then analyzed with Spark and Jupyter notebooks, storing forecasts and meta back into Prometheus via a custom exporter. Uses Prophet and Fourier models retrained on configurable windows (example: hourly retrain), applies an accumulator to reduce false positives, and reports upper/lower bounds plus anomaly flags. Prototype handles about 360,000 metrics/hour and is packaged as a container with CI setup and GitHub notebooks and exporter code. Prometheus → Thanos/Influx → object storage
AI-written summary · about
chapters (11)
- 0:00Start
- 1:32Ai Center of Excellence
- 3:15The Open Data App
- 7:06What Is Prometheus
- 8:46Targets
- 16:11Prometheus Metrics Tab Metric Types
- 17:04Histogram
- 21:19Finding Anomalies
- 25:01Prometheus Deploy
- 25:48Prometheus
- 30:16Fourier Analysis
related talks
37mAIOps: Anomaly Detection with Prometheus and Istio
37mAIOps: Anomaly Detection with Prometheus and Istio
31mHandling Billions of Metrics with Prometheus and Thanos
27mObservability with Prometheus and beyond
36mObservability with Prometheus and Beyond: Leveraging Cloud-native Technologies Outside of Cloud
35mAnomaly Detection for Cloud Native Storage
10mUpstream-First, High Scale Prometheus Ecosystem
02h 43mPrometheus from A to Z