Your question very field dependent. In my social sciences field, literature reviews 1) answer some kind of open question in the field, just like an empirical paper does and 2) are, these days, usually systematic reviews.
You'll hear a lot of researchers talk about the right "time" for a literature review. They may say it's too early for a lit review in the field, or that it's the "right time." What that usually means is that there's enough collected empirical research in the field to do the review but ALSO that there's some sort of open question or idea that the literature review contributes to. Turn it back on yourself: if you were looking for papers in your field, what kind of literature review would you want to read? Methodological ones are good; there's a well-cited one in my field that reviews the empirical work in my area using a specific method that I use, and it's very useful in conceptualizing and writing my own papers.
More theoretical ones can be useful, too, if you're trying to answer a question: What really IS this X that people in the field have been talking about in 27 different ways - can we come up with a unifying concept or underlying theme that makes it easier to talk about that X? (I think that's the first idea that you highlighted in your question.) There are about a million papers on whether Y affects Z; have we come to some sort of conclusion as a discipline about whether and how Y really does affect Z? In my field, the people who write these papers are usually senior scholars (or a team that usually has a senior, established person on it). Some journals don't accept them unless they specifically invite the scholar to write one, or unless the scholar contacts the journal ahead of time with an outline or plan for what they want to write.
Basically, journals don't want a random summary or even a simple synthesis of several papers. They want a review that's going to push the field forward in some way. Most dissertation literature reviews don't do that - but if you can write your review in such a way that it does (or at least lay the foundation so that when you go back to it later, it's easy to transform it into one that does) you'll have an easier time getting it published.
The other thing is that in my field, these days, most journals want systematic reviews. It's a very specific way of doing a literature view that makes it akin in format to an empirical paper: with bounded year dates, specific search terms, and a systematic way of deciding whether to include or exclude papers (hence the name "systematic review"). It's a way of trying to ensure that authors don't only include papers that support their views. Many times systematic reviews also include a meta-analysis; I suspect that they are more likely to get published if they do.