I have found a very good overview of a topic and I want to cite it in the introduction of my paper. The writers are very well known people in the field. Is it then okay to cite the paper that is not published and is only available in arXiv?
There is nothing wrong with citing paper, reviews, etc., which have appeared on the arxiv (and only on the arxiv). Go ahead and cite the paper.
(If your paper gets accepted at a journal, they might ask you to update in case the paper has been published in a journal in the meantime, but in the fields where the arxiv is commonly used, journals and referees will have no issues with references to the arxiv.)
I recommend using "peer-reviewed" rather than "published" for a paper on arXiv (or otherwise on the internet), because such papers are literally published, in the sense of being publicly available (even more than those behind pay-walls). The distinction that may matter is peer-reviewed (or not).
The question of whether peer-reviewed papers are more reliable than arXiv papers by well-known, reliable, experts is a separate question.
But my point, somewhat tangential to the literal question, is that "things on the internet" are published. They are cite-able. The question of their relative reliability is subtler. Peer review is no guarantee of correctness...
So if you use something from arXiv, cite it. These days, in math, for example, everything appears on arXiv or on peoples' homepages a long time before it is "peer-reviewed", and the latter may be "never"... simply because in some peoples' perceptions that "peer review" does not add value, but is fairly expensive in time and trouble.
If you use the work and you found it only on arXiv then you must cite it there (or elsewhere) to avoid plagiarism.
But, the publication process takes a long time. I suggest that you cite what you have found, especially if the paper isn't formally published yet and watch for it to appear. If it is ever going to appear it may well show up before your paper is finalized.
But failure to cite is a serious matter. But a more permanent citation is preferred over arXiv. You don't need to wait for it to be finalized to cite it.
You can cite everything that you want, including arXiv contributions. However, you might want to look for alternatives. In case of review/overview articles you often will be able to find some and use them instead without risking plagiarism. In academia, peer review is the standard way to make sure a paper meets the necessary quality criteria, so peer reviewed publications are more trusted sources.