I don't have accurate enough knowledge to be sure that these are all real problems. It seems like they'll be problems, but maybe there are prevention mechanics in place that I didn't manage to notice. Remember that some of these concerns are based on how players will behave given the game systems, which a few hours on a test server won't reliably demonstrate.
1. Contention on killing undead in Delera's graveyard. In games with a lot of open-access PVE content (meaning another player can join your side in battle without a specific invitation), the designers wind up doing a lot of careful work to administer how credit is assigned for killing monsters. Take a look at the "tap" system in LOTRO to see how much complexity can go into balancing the needs of spontaneous assistance against interfering and sapping credit. It appears the Mabar event may face this problem in two ways:
A. On the back end, if someone is there battling an undead and I come up behind it and attack, am I helping him to victory? Or harming his chances of collectibles to enrich myself? It seems mostly like it'll be the latter.
B. On the front end, if a wizard zaps 15 undead with Ball Lightning. does that mean he owns them and other players would be trespassing if they attacked?
2. Instance-hopping to find fresh undead victims. Is that effective? Should it be?
3. Lowbie support. How much can a lower-level character participate in undead hunting around Delera's? If players do that, will it turn out to be a waste of their time they could've spent levelling? (I noticed that some harbor lowbies got upset by the Risia event, because it lured them into investing time trying to get something fun and productive out of it, when they would've been more entertained sticking to Waterworks and Braza's Farm)
4. Spectral Dragon encounter concept. Splitting into 4 groups which battle a dragon one at a time as it moves between rooms... that seems less fun than alternative layouts where the majority of the players can see the dragon for the majority of the time. Only 25% can be there when it's killed, and if you fail there's only a 25% chance you were in position to see what went wrong. (Compare against pillar-killing in VON6. That also has the players divided up and dependent on each other, but at least it allows the possibility for someone to run from one platform to another to rescue a group having trouble)
5. Spectral Dragon key cost. Someone who pays cash money for an ethereal key will be additionally disappointed when his key is wasted because some other player (whom he didn't pick or even interact with) causes the dragon battle to fail.
6. Spectral Dragon group selection. The random FCFS way players within a level band are placed into the Spectral summoning rooms can be troublesome. Suppose there are only 3 other guys attempting it in your level band... you don't know that until you've spent your ethereal key, so then what? Wait there for 1 hour, 2 hours? Blocking players outside from interacting with the traders?
6. Spectral Dragon group size. Allowing 16+ people (or whatever) breaks and stresses many things. Aside from stuff like client loading and game balance of AOE vs single-target abilities, it also means our party list and chat systems aren't fully functional.
7. Named item balance. Mostly not a big problem, but just taking a look at it:
A. Handwraps seem highly superior to other available undead-beater weapons, especially as they're pretty easy to acquire. Seriously makes me wish that non-Monks could make unarmed attacks at their full melee speed (which is what happens in the D&D rules, particularly since they can elect to learn Improved Unarmed as a feat). Monks were already relatively more effective against undead, and this magnifies it.
B. Cloak less of a problem, but it is pretty superior to the conceptually-similar cloak from the Abbot raid, which is far harder to acquire. The Abbot one has the advantage of providing Sneak Attack, but most people just use Tharnes. It also competes well with sources of DR, as the Good bypass is quite meaningless.
C. Robe even less of a problem, but once again it's superior to named items that are way harder to obtain.
I'll repeat a suggestion from earlier:
Change the Spectral Dragon summoning place to a standard instanced 12-player raid, with disticnt versions for level 5, 10, 15, 19, 20 (or however it is divided now). Players can only enter if they have the key and Joe has gotten his items to temporarily open the quest, and they'll manually assemble a team for the mission just like they group up for any other raid.