Your exposure is 10 stops below "sunny 16" average direct sunlight, and at f/1.8 your depth of field is quite thin.
Autofocus is an imperfect technology and yes, I think you are pushing its limits. You are getting an 83% success rate .
Nikon specifies that the camera will autofocus between eV -1 and 19. Your image is about eV 5, suggesting typical interior home artificial lighting. That is not a lot of light.
In principle, you are 6 stops above the minimum eV, but Nikon does not specify or guarantee 100% consistency at any of those eV levels so there is no way to pin down the expected performance based on specs.
In my experience, if I got your results I would consider it expected. And in most shooting there is a large element of human error. For that reason alone I tend to shoot multiple images in difficult conditions, including the light levels you have there.
I don't think your "out of focus" examples are all that bad, and about what *I* would expect at those light levels and depth of field. The variance in focus is far less than the 2-3mm differences in the various targets, suggesting any error is some fraction of 1mm.
Since you bought the camera used and don't know how it was used (or abused), it is *possible* that there is a little dust on the internal AF sensors that might be removed, very marginally increasing performance. But that is just a (remote?) possibility; I would not go through that expense based on these images.
If you have some critical use that requires better consistency you might want to take the long odds on that and get it cleaned just to be sure.
In discussions here regarding focus tuning, the problem of consistency is often discussed, even with brand new cameras. It is the state of the technology.
What is your working distance? (distance from the target to the camera film plane mark (roughly middle of the body, not measured from the lens front).