Model Monitoring and Drift Detection overview

Continuously monitors deployed model performance for data drift, concept drift, and prediction quality degradation — alerting on threshold violations and triggering retraining workflows. Excludes initial model training.