Hacker News has been up and running for like 10 years now, isn’t it? Instead of hypothetical question, I guess we have actual data on the uptime for such question.
I think redundancy is much easier if the thing being made redundant is extremely simple? Imagine having to manage 50 different services (and sometimes servers) as opposed to 1 machine.
I guess the advantage of not relying on your site for profit is that you can afford to not worry about a bit of down time so much...