I think, to some degree, we need to look at it from a % chance of it being effective.
I'd think a lvl 10 AA probably has a first attack AB in the ballpark of +30, and most of the foes that character is going to legitimately encounter are going to be high-to-uber range
Just keeping the numbers simple, but you can tinker with them however you like to make your own example:
If the foe has an AC of 40, then that's a 50% chance to hit
Then the Death Arrow then has to beat the Fortitude save, keeping in mind that anything with a Fort Save of +19 or higher is just going to be immune, since the 1 roll is not an automatic failure for saving throws. I don't know what the average Fort save for a high-end mob is off the top of my head, but let's say it's +15
so that means it's got a 20% chance of success.
That means the Death Arrow attack has a 0.5 x 0.20 = 10% chance of successfully instantly slaying it's target.
Throw on concealment and that's another 50% haircut to the success chance, bringing it down to 5%.
You can manipulate these numbers however you like using whatever attributes/skills/feats/spells you like, but it seems to me the chance of the Death Attack succeeding when fighting a fair-game opponent is going to be generally below 15%.
We tested this in a friendly spar last night and it worked out the chance of success would have been a frustratingly low ~2%:
90% chance to hit x 50% from concealment x 5% chance to beat the Fort Save = 2.25%, or 4.5% without the concealment.
Admittedly, characters who are threatened by a lvl 10 AA are probably overall going to have a lower % chance of this successfully affecting them since they tend to be better equipped/prepared than your standard dungeon mob, but even as the potential spar/PvP victim, that struck me as very low-threat for an attack type from a highly-trained PrC that you can only use once between rests. It relegates what is a Capstone PrC Ability to the realm of "Hail Mary, Last-Resort, Desperation Attacks."
I also don't get the sense there are legions of level 10 AAs around that pose a threat in this manner, either, so I tend to agree with some of the other posts above where there could be some variable factor that might help raise the DC vs Fort Save of this attack into at least the mid 20's.
Just my two cents.