As a general rule, unless given an explicit explanation for discontinuity on screen, it should be the explanation of last resort.

The problem is that as an explanation it can be used for everything. Consider any shot production error that might happen. Actually let's use one of my favorite TNG episodes for discontinuity: Parallels. In the final scene of Parallels, there is a continuity error, where a bow switches sides.

Does this mean we should perform an inception style deep dive and say perhaps Worf is still jumping universes? Could we use this to, in fact, explain ANY minor production error?

I mean we could. But that's probably not what's intended by the authors.

For example I am very much a fan of the idea that early in TNG's run the Ferengi still valued gold and later on they do not, and this matches up with better and better replicator technology eventually being able to create gold at scale. But also, maybe it's just temporal discontinuity.

Can we reconcile Picard's relationship with his mother with what little we see from TNG and what we see in ST: Picard? This can be a fun exercise. But we can also say "Eh not the same Picard."

The idea that Khan is destined to happen is a heads on explanation for the intractable problem of Star Trek is rerooting its history into our modern history. Star Trek is, after all, a vision of our future and that vision has changed from the 1960s. This is a change designed to add some meaning to the show.

On the other hand, if "time pushback" is used to explain anything and everything on the show, it runs the risk of becoming flat out meaningless.

So when would I consider it an acceptable explanation? Whenever it's given as the explicit explanation, or maybe if there's a very clear connection.