Have a look at this post, it might be the solution to your case as well: https://community.mendix.com/link/spaces/microflows/questions/131234 without knowing the details of course.
You might want to take a closer look at event handlers (Before/After Commit) on the entities involved. Even if it looks like concurrency isn’t the root cause, event handlers can still cause deadlocks when multiple updates or cascaded commits occur inside a microflow chain. Try temporarily disabling them or adding detailed logging around those handlers to see whether they are contributing. This kind of debugging approach often helps pinpoint unexpected commit sequences that lead to deadlocks.