To expand of this a bit: Failure is part of production. So having stuff failing is part of your infrastructure work.
At the same time, pratice shows that a lot of "test setups" turn into production without ever being setup from scratch, so treating everything like production really helps.
This doesn't mean you shouldn't have any test setups, but unless you have a separate production environment, there is only production.
This is my personal microblog. It's filled with my fun, joy and silliness.