mlops_integrations

MLOps Integration Examples

DriftAudit returns the standard AuditReport, so monitoring jobs can publish JSON and HTML artifacts, use .rai-audit/history, or forward summary metrics to an observability platform.

MLflow

Call log_drift_report(report) from mlflow_tracking.py inside an active MLflow run:

import mlflow
from mlflow_tracking import log_drift_report

with mlflow.start_run(run_name="weekly-drift-monitor"):
    report = audit.run()
    log_drift_report(report)

Airflow

airflow_dag.py shows a weekly DAG that runs the batch monitoring script. In a production DAG, replace the example command with the job that loads your reference window and current batch from your feature store.