This is very simple: You are the editor. This is not a problematic practice but just part and parcel of the review process.
If you are the editor, you get to make the decision. (I assume that there is a way to handle complaints about your decision, as most journals / professional organizations provide that.) A referee is obliged to check whether references are suitable. You as the editor in charge selected the referees because they knew the field, so it should be relatively common that the referee also has published in the field. Objectively, there will be cases where a referee's publication should be cited and maybe discussed into the revised version of the paper. Subjectively, a referee will sometimes err and occasionally sin by demanding that one of their publications be discussed.
(Also, there are predatory journals who will ask a paper to cite other papers from the same editorial group, but this is not our topic.) This is where your job as editor comes in. You need to make a decision:
You can tell the author that they might want to consider the request but that acceptance will not depend on their decision.
You can tell the author that they definitely should include the reference and discuss the connection. You can also invite them to send you a message discussing how they see it.
You can tell the author that the paper is rejected because it duplicates results that the referee already published.
And there are of course more possible outcomes.
However, you are the editor and you have to make the final decision. If you feel uncomfortable or need help, there should be other editors or maybe a senior editor to ask for advice. You as editor are the guardian of scientific integrity.
You would be a fool if you invite someone to review a paper and then do not listen to what they say. But you are not the reviewer's peon.
If you feel that the reviewer is revealing their identity, you can redact what they are saying or play this back to the reviewer if the editorial system does not allow for it. However, just because a review says that the results of X and colleagues should be referenced does not mean that X or a colleague of X is a reviewer.