DevConf.CZ 2020

Monitoring your applications with Prometheus and figuring out relevant metrics to alert on can be a tough task. What if you had an automated AI based technique to help you identify these metrics? Introducing the Prometheus Anomaly Detection framework! In this workshop we will walk through all the tools required to setup your own anomaly detection framework for prometheus metrics. We will see how to:
1. Setup a sample application to generate metrics
2. Configure Prometheus to collect the metrics
3. Use a python library to get metrics into a suitable format
4. Train machine learning models to perform time series forecasting
5. Use Grafana to create insightful dashboards and setup alerts