Machine Learning Model Monitoring in Production: Monitoring Methods and Tools

One of the key problem related to Machine Learning model in production is model decay, meaning the performance of the model degrades after deployment, and a poor solutions is provided to the business problem. Monitoring is crucial in the process, as it helps detecting when model decay happens by keeping an eye on the monitoring metrics. In this nugget, we address the question:How do we know if monitoring metrics are indicating an anomaly?