Your criteria must be faulty, it shouldn’t be tested in production if it wasn’t already production ready.

I think similarly when considering if I’m willing to deploy something though; if there isn’t a long running sample why would I trust it?