An exceedingly complicated and controversial question! Some have argued that you only need epistemology, or even that epistemology is all you can get; you can only know what you can know, so you might as well confine your attention to the knowable, and not worry whether there might be things that are which are not knowable. Others claim that it's obvious that whether things exist or not surely doesn't depend on whether they're known, and it's even less likely that it could depend on such a suspicious, hypothetical property as knowability. Of course, the latter view doesn't entail that one should favor ontology over epistomology, but trying to balance both introduces very difficult problems of how to tie the two together, so it is fairly common to take one as primary and use it to settle questions about the other.

One might wonder what practical consequences one choice or the other might have, and here again there is much controversy. The pro-ontology faction claims that emphasizing epistemology encourages subjectivism and relativism and weakens our grasp on reality. The pro-epistemology faction replies that emphasizing ontology is exactly as relative (or non-relative) as emphasizing epistemology, it's just that when ontology is emphasized, biases are hidden because the focus is turned away from questions of how actual humans arrive at their ontological conclusions.

Personally, I am tentatively on the side of the epistemologists, but it seems to me that details matter a great deal, and there are far too many details to discuss in a comment (indeed, a book is likely insufficient).