Monitoring in practice.

Setup is a decision, not a project: pin the metrics that matter, and monitoring does the rest, re-checking each as new cases land. The daily experience is reading what surfaced, not remembering to check.

Pin what matters

Any metric the platform computes (e.g., standard endpoints or your custom algebra) can be pinned to a cohort's watch list. Scope decides the question: the same metric pinned per machine, per TPS, or per technique watches for different failure modes.

The monitoring run

As new cases accrue, Radial recomputes the watched metrics, updates the control charts, adjusts for case mix, and decides whether anything deserves your attention. When nothing has shifted, nothing surfaces, and that silence is the feature working.

When something surfaces

A shifted metric arrives as a control chart on your Home screen. The review is a physicist's, not a statistician's: toggle raw against case-mix-adjusted to see whether the patients or the process changed, narrow the scope, and drill from any point to its case. When several plans share a sim date, you pick which to open. Radial raises the flag; what it means is your call.