I think this depends entirely on what you are indexing. From my experience with some 100 ish scrapers for news sites a few would break literally every day.

And the only thing we really wanted was article title and date.