I will try to give an approach to assessing this position. My aim is not to reach a conclusion but to illustrate a process. It is a typical error to try and assess a position by looking at where the pieces and pawns currently stand, instead of asking where they can go to.
OP does somewhat acknowledge this by saying "about the knight, with the current pawn structure I don't see an obvious better place for it." Well, how about f4? We get off to an immediate start with ..Nh5, threatening to win a piece. If then Bh2, then ....g5 (the pawn structure is not permanent) the Knight is going to f4 and if it is exchanged off we retake with the g-Pawn
which will give us an open file for attack that outweighs the doubled Pawns. Oh, but there is a downside to ..g5. White will play Ne3 (that Knight was not so badly placed- I feel sure it went to d1 with this in mind!) and now White too will have a well placed Knight unless Black trades off that beautiful Bishop. If he does so, then if White is afraid of doubled pawns he will recapture with the Queen, but much better would be to retake with the Pawn, kicking the knight from its outpost. So we decide to keep our Bishop. and we have the line of play ..Nh5, Bh2 g5, Ne3 Nf4, Nf5 and now where does the Queen go?? Both e6 and f6 would protect the h-pawn, but e6 has the extra merit of combining nicely with a ..g4 push.
At this point I find myself liking ..Nh5 but suddenly notice that instead of Ne3 White might try g3, which is so ugly that it does not easily come to mind, but does defeat Blacks aims. If the Black Rook had been on f8 I could keep up the momentum with ..f5 So then I start to consider playing ..Rf8 at some stage, or even playing (in the position after ..Nh5, Bh2 g4, g3 f5 as a temporary sacrifice. But after Ne3 I find myself unable to do better than to get that Pawn back with ..Bxe3. My assessment now is that both positions have been damaged, and perhaps more so for Black.
So that did not work well, but I was trying, as I said, to illustrate a process. The engine recommendation of ..d5 does not seriously weaken the e-pawn. In fact after ..d5, exd Nxd5, it is the WPd3 that starts to look vulnerable. If White protects the e-pawn with, say, Qc2 then Black can play..Rad8 followed by doubling on the d-file. The move ..d5 is probably what should be played.
I will try to outline a procedure that might be followed, beginning with these questions.
Is the position stable? In particular is the Pawn structure stable? Which pieces are likely to still be on their present squares in five moves time? In ten moves time?
If a piece is likely to move, do I want to exchange it off? If not, what square would I like it to be on? What useful Pawn moves do I still have?
Can I effect the exchange? Can I route the piece or pawn to its preferred post? If not, can I remove the obstacles?
This is harder! The same questions from the opponents viewpoint!
Select "candidate moves" on the basis of these answers. Analyse the tactical consequences of each as far as you are able.
Entire books have been written on this topic. I doubt that anything I may say here can truly answer your question, but one thing I feel sure of is that you need to base your evaluation on more than the static features of the position. And although you are correct to look for negatives, like the possible weak e-pawn after ..d5, remember that unless your opponent is a complete beginner you will not not be able to give weaknesses to him without accepting some of your own. Good luck!