because you define thresholds for event classification.. the difference between 10024 customers having failure and 10100 customers having failure is not the question, right? when many hundreds of thousands of customers are failing at once, is that really very difficult to determine?
secondly, there are financial, management and information security pressures to NOT REPORT reality to the public. This happens VERY OFTEN in real business. In fact, that is why legal enforcement actions and real consequences are crucial versus Big Business.
In a cloud, massive outages are the rare events, and technically easy to report.
Small outages happen all the time, and are difficult to report accurately.
I think AWS has pivoted to trying to report status in each individual customer's support portal, so that they can give a dashboard that reports the state of the cloud from that customer's perspective. That way a rack down that only affects a few customers is only reported to those customers, and the dashboard doesn't have to always be red for everyone (or green for everyone, even those affected).
secondly, there are financial, management and information security pressures to NOT REPORT reality to the public. This happens VERY OFTEN in real business. In fact, that is why legal enforcement actions and real consequences are crucial versus Big Business.