I'm playing a (turn based) game in an opening line I've played before; in fact, a few moves ago we reached the point where I had played it before in 1 game.

I did some analysis after that game ended (concluding I was better at one point but later on lost the thread), and I used engines then as well, after the game. But I also noticed just now that I hadn't actually used chess.com's Computer Analysis feature on that game.

Is it cheating to order computer analysis on that old game now?

Considering you are triggered because you have that same position again in a current game,.. I would not do it.

I don't know if it is official considered cheating. But it certainly is not sporting behaviour

It feels like it's bordering on unethical due to the way you came up with the idea of analysing the game, but under other circumstances you would be fine. 

I don't think anyone can accuse you of cheating though (you'd be clear under even strict rules) -- you have the right to analyze your games, after all, and it's common that people play the same openings over and over again.  There's no rule -- in ethics or elsewhere -- that says a correspondence player needs to analyze their own games promptly after the completion of a game and before another one has started, or else not do it for a while: that would be discouraging learning.

There are other such corner cases in correspondence play due to the length of a game and the non-simultaneous completion of other ongoing games.  In all of those cases, I believe you are clear.

I must say I did order the analysis first, and then I realized there might be an issue.

We're still in main-line theory. When it's done I'm only going to look at the last few moves of the game, where I seem to recall I had a way to save my game but blundered, but I can't remember what the saving move was. In my view that's OK.

I think the ethical way is not to use any computer analysis of this old game for your opening play in your current new game. Next time you run into the variation, you can use it.

That's what I would do anyway

I don get it.

What's the difference of ordering computer analys here and doing a computer analys with your own chess engine?

I analys all my games afterwards with a chess engine (Rybka, Fritz etc.)

I also get information how to play a opening from chess databases, books etc. etc.

I havent had this issue yet, but I have a few chess opening books where I have jotted down moves recommended by fritz, usually where the book didnt consider a move I was concerned about.  Sort of strikes me as a gray area.   

I was a little confused because in general not only post mortems are fine, but engine use is also perfectly normal; except of course when it's engine use in positions closely related to the position you have a in a current chess.com game.

So although I used engines on that previous game after it finished a month ago or so, I wouldn't do so when I had an ongoing game in the same line. But the computer analysis is a chess.com feature... didn't feel the same.

There's a huge gray area anyway. Say I'm computer-analyzing some idea for White in the Budapest Gambit with 4.Bf4 g5 (I am, and I do plan to use the results in the next chess.com game where I play that variation -- that's perfectly OK, it's basically the same as using a book). When must I stop doing that? When someone plays 1.d4 Nf6? 2.c4 e5? 3.dxe5 Ng4 4.Bf4 g5? Sometimes they take ages to answer even in the opening...

It's 2 different games. I don't see the problem. If you chose to analyse the first game after the fact and figured out where you went wrong then I don't call that cheating, it's learning. I'm not sure what you mean about the analysis feature on this site if you've already looked at it on your own software then why bother?. If I understand correctly it's probably best to leave it until you finish current game just to be safe, but I still wouldn't consider it cheating. Coincidental.  Lucky maybe.  

During my younger days when I used to go for Chess class (practice sesion) every Tuesday evening, one of my friends could easily memorize the first 10 moves of a major opening moves eg Grunfeld, Sicilian, English, French, etc. He could even play blind-folded.

During those days, he was one of the few chess players whose rating reached 2000. (unlike nowadays where an ordinary chess enthusiast, with a litle bit of dedication, could easily reaches 2000).

I've learned most of my chess repertoires through his games. His famous comment, "don't ever forget to analyse your game after you are finished".

Basically, when we are doing the analysis, we are trying to spot where the mistake was - a good chess player always learns from their mistake. Should an opportunity arises for them to play the same line, they will repeat the same line all over again and only diverted where they had last commited the mistake. (This is what an analysis is all about). In the end, he/she will be the master of that line. (Anyway, this is how certain variations are named after the Grandmasters).

A good chess player will always take some time to analyse their games, be it a winning game or a losing game. In chess, there is no "perfect" move. To me, if you are trying to follow the same line all over again, it can't be considered as cheating. You are just trying to improve on your line (to improve your chess repertoires).