It depends on whether the party HAS to get through the door or not, doesn't it? One of the reasons behind "fail forward" is specifically to teach GMs that there is a different way to deal with a "single point of failure" problem. That is, sometimes you're going to only realize something was a single point of failure too late to directly address it--or you're improvising and didn't think that far ahead, or you truly want this to be a single point of failure because that creates tension, or whatever else. But a single point of failure where the only result of failure is "the game grinds to a halt because nothing happens nor can happen" is pretty blatantly a bad thing.
Fail forward, as a rule of thumb, means that even if the party gets struck by such a thing, the pace of the experience and the enjoyment of play don't get drained away as the party sits there, waiting for one of their schemes to finally, finally, FINALLY open the stupid friggin' door.