I'm 2100 in online blitz so we have a similar rating. I'll give you the same advice that Judit Polgar gave to me. Study middlegames and basic endgames. And tactics. I think that Dvoretsky's Endgame Manual might be difficult. It's a little too in depth. It's hard to apply the knowledge that you glean there to your games. For the amount of time you invest in it, you get very little application.
Let's explore this in further detail. There are three components to Judit Polgar's advice: (1) middlegames (2) basic endgames and (3) tactics. It's important that she left out openings, because at the 2000 level it's more important to know plans and ideas than to memorize sequences of moves. It's also important that she said "basic endgames" and not "thoroughly know all the theoretical endgames", because basic endgames have enormous application -- such as opposite colored bishops, a- and h-pawns, knowing to put your king in front of your pawn and your rook behind your pawn -- whereas knowing the theoretical diagrams in Dvoretsky's book has little application unless you're a grandmaster. Why? Well, even if an endgame is theoretically drawn, at the 2000 level your opponent will often make mistakes which render the theoretical result irrelevant, since what's really relevant here is whether you exploit their mistakes and use that to win.
Instead of Dvoretsky, I would recommend looking at annotated games (e.g. books, Chessbase, Chess.com) and watching top players play. This will help you absorb plans and ideas that you can then use in your own games. If you have a good plan during the middlegame then on the 2000 level you often have an edge against your opponent. Forming a knowledge base of plans, ideas, motifs is a lot like building a vocabulary. In order to build a good vocabulary it helps to be a voracious reader. Well, by analogy, you want to be a voracious chess observer. In addition to reading annotated games, you can watch the world's best players play blitz on chess.com. Why blitz? Well, Magnus Carlsen recently said that blitz and rapid are in a sense a purer form of chess because it takes the element of preparation out of it. It's easier to understand a grandmaster's ideas in blitz or rapid because the barrier of opening preparation, to which you are not privy, is removed to a certain extent.
If there are positions where you feel uncomfortable, such as isolated queen pawn positions or even the endgame itself, then it helps to play those positions at every opportunity with willingness and enthusiasm, so that way you can get a feel for them and improve. I think it was like eight or nine years ago that I was afraid of the endgame, and preferred to win by attacking my opponent in the middle game to get a decisive material advantage or a mate. To improve this hole in my game, I just decided to play endgames with (1) willingness and (2) enthusiasm, often trading off pieces just for the sake of entering the endgame, especially when it was equal. Then I started winning equal endgames, and I found it enjoyable. What improved my endgame was not Dvoretsky but a basic endgame knowledge and lots of practice - and, of course, willingness and enthusiasm for the endgame.
But this last bit of advice is important. If you watch grandmasters play blitz on chess.com, you'll notice that even 2700s and 2800s miss basic tactics that afford their opponent the opportunity to win. That's why it's so important to learn tactics. You can use them to win at almost any level. The number of tactics available to you in a position depends on how well your pawns and pieces are placed. The better position you have, the more tactics you likely have. In order to improve tactically, you also have to improve positionally. That's where studying the middle game helps. When Polgar gave me that advice, she said something like "spend 70% of your time on the middle game". Because if you have good plans and ideas in the middle game, then chances are you'll reach positions where you have winning tactics. There's a big disconnect where some players place their pieces well but are then unable to find the winning tactics. I think this comes with experience. The better chess shape you're in, and the more experience you have, the more easily you can convert advantages, and know when to convert advantages. It's a skill in itself. Some times you have a strong knight on d5, but then you have to give it up in order to break up black's pawns on the kingside. Some players would be unwilling to give up this knight and never make any progress in their attack.
At the 2000 level, so many games are decided by not seeing one or two moves ahead. In other words, tactics. In order to get those tactics you need to have a good position. Hence Polgar's emphasis on middle games. When you go online, don't forget to watch top players play. It's something that I only started recently, and I wish I had started a lot sooner!
