That's really interesting. We didn't get any user reports before our canaries fired, either. Now I have to think about what might explain the difference between your systems and ours. We're monitoring API Gateway health (more or less), because that's what we care about in this part of our infrastructure.