Informal fallacies are judgment calls about problems in the structure of an argument. Their applicability is often debatable. (Formal fallacies like affirming the consequent are different in that they show the actual invalidity of an argument).
The fallacy of tu quoque is to fallaciously deny an argument because the source of the argument does not follow their own advice. But whether it is fallacious to point out that they don't follow their own advice depends on an interesting problem that we see described as far back as Aristotle.
Aristotle talks about akrasia, which we often translate as weakness of the will. It refers to a problem in the "practical syllogism". The practical syllogism is the idea that if I know what I ought to do, then I will do it. It turns out we often fail despite appearing to have full knowledge of what we ought to do.
Giving advice about, say, weight is an example where it seems fully plausible that a person giving it may rightly know what they are talking about even if they fail to follow it. Or to put it another way, if the leading scientist on obesity also happens to be fat that does not make what he's saying fallacious. On the contrary, we'd be guilty of a tu quoque to dismiss his knowledge in that arena. There is something bizarre seeming about following his advice on that point, but it's not fallacious to do so.
Consider for instance another informal fallacy: argument from authority. If we listen to the world's leading expert on obesity and do something about our weight because she says it, then that's an argument from authority. But that doesn't mean it's fallacious to follow her advice. Why? Because informal fallacies require judgments about the relevancy of her expertise.
To sum up, informal fallacies are patterns that are generally errant due to a problematic feature. But that feature is not always problematic, and we have to make a judgment about whether it is or not.
I don't generally take weight loss advice from really fat people, because (all other things being equal) it is difficult to accept they are experts on how to accomplish weight loss. But there may be exceptional reasons that make it so it's not fallacious. Conversely, there are contexts where tu quoque is perfectly appropriate (e.g., imagine one smoker telling another that it's easy to quit when the other smoker can't quit themselves. It's perfectly appropriate proof that their claim is wrong that they themselves empirically invalidate it).
