Very high server uptime is a sign of dysfunction. Every server I've seen with a multi-year uptime was some hideously outdated single-point of failure that nobody had the courage to touch.

That's part of our civil defence strategy. Do you think it's an accident that everyone ends up with so many spare kettle leads?