The 3 phase plan does look good. Possibly a similar plan could happen for slightly bumping up the healing encounters need too.

I can imagine some more simpler ones like the standard AoE pulses and that bleed you were talking being inserted on the first phase. The healing absorb (Aetherblight) could be the 2nd step. We'll probably have to keep the aetherblight pretty low to ease the other players in if they like it. The numbers we discussed for a full 1 -4 health bars worth of stopping the healing will probably end up being Savage mode testing. I am trying to predict how many would try to instinctively use Esuna on it, but I may or may not be underestimating what would happen. If it's like the other debuffs with that underline on it or so, it might be a good enough indicator they can dispel it. Phase 3 would probably be fine tuning some numbers for the encounter damage, aetherblight levels and the self heal levels of the tanks and DPS.

I can see your point for the abilities inevitably having some form of similarities no matter how we approach the spell / skill changes. With 21 classes, it's probably impossible to make everything 100% different without vastly breaking the balancing of the damage and healing dealt. The cooldowns you compared definitely cement the 2 minute meta to how it is with Critical Strike / Direct Hit usually being a top stat all the time. It does spark a question if a class meant to do just filler DPS could work. It would likely need multiple traits to fine tune the filler to be a higher average than others at end game and similar to others at 50.