Correct, but I prefer to understand the actual downtime experienced by people over the years (eg data centre issues, network issues, unplanned server maintenance) as that data seems hard to find centrally.