Very true - the complexity of the system also comes into the equation when things go wrong and it takes real people to figure out why an outage is happening. Highly complex systems imply longer debugging time, and at a certain point a theoretically lower up time can give you higher practical up time just because engineers can actually understand and debug it.
