I've been running on Cloudflare at one of my locations for about 1.5 years. There has been some down-time, but it's usually been on a local cache and not a national outage like this one. They were pretty responsive -- the total downtime was somewhere around 32 minutes.
It took me three minutes to isolate my problem to them. I just added some non-Cloudflare DNS servers to my resolver. It's easy to look at the incredible redundancy that's built-in to their system and be confident, but adding a higher layer of redundancy to an already redundant system is never a bad idea.