How do you implement alerting and monitoring for a pipeline that fails silently due to upstream data drift?
