I've been singing the praise of AWS Route53 for a long time, they up and running. I can't believe major multi-million dollar companies (Twitter, GitHub, Soundcloud, Pagerduty) would not run a mix of multiple DNS providers.
Also what is happening is a cascade effect, where a 3rd party being down effects others.
One of the reasons why Route53 is good is because they give different nameservers to each hosted zone - unless you choose to use a branded record-set.
I've seen them be hit by dDos attacks in the past, but never had any significant impact.
(I wrap Route53 and handle storing DNS records in a git repository over at https://dns-api.com/ Adding support for other backends is my current priority to allow more redundancy.)
Also what is happening is a cascade effect, where a 3rd party being down effects others.