For 99% of people Node would be way more than enough.
Can you back this up?
I’ve been in IT for over 30 years. It’s a common understanding that server applications tend to be large and sophisticated, and that for mission-critical enterprise projects, reliability is a very, very crucial metric. Google have said as much with respect to their internal use of Go. What kind of backing are you looking for? Statistical studies?
As for debugging Node, this InfoWorld article gives an example of how Joyent dealt with the problems of debugging Node. It’s usually an onerous process compared to more conventional server applications written in sane programming languages.