Spoiler: I haven't run this adventure. I am running Barrowmaze, which I will review once finished, but thing about Greg Gillespie's dungeons is that often the random tables get heavy heavy use. And often the random tables both revel the plot, and are the most dangerous encounters. This is key OSR. Pathfinder has always taken a different view, that dungeons are prepopulated, every room is explained, and combat is likely to be 'fair'.
This book tries to use a method similar to Gillespies, and apply it to Pathfinder. Does it work? Sort of? But first some related comments:
Firstly I have the print on demand volume. It's big. Like biggest book on my shelf big. 837 thick pages. I'm legitimately worried the spine will not hold up to age or play. Honestly I think the book would have been better being in two volumes.
As with Dark Obelisk 1 the author has tried to colour code everything. But they've used some very solid hues of colour and instead of helping, there is just too much different coloured items it looks a mess. It's a great idea, but the colour scheme itself needs to be worked on.
The maps (in general) are excellent. There is a large overview map of each area, then smaller more detailed maps of each section. Some maps (such as page 249) don't have a grid when they should. Unlike the previous book there are not player facing maps, which is a huge shame, as printing out handouts from the book shows the secret doors. And running this dungeon I'd want to print out handouts of each section (ideally unnumbered) as there is enough detail there for players to really run with.
My next comment is that the random encounters are not actually in the book. They are in a separate pdf. As are the random treasures, and all the core random tables. Not cool. It doesn't even look as if there is an option to buy this 92 page pdf in print. So if you are a printophile (like I am) this puts you out of luck.
Now we come to the default check time for wandering monsters - 5 minutes. There is an encounter introduction table - but it only focuses on combat, the size of the map the monster(s) arrive from, and who is surprised during combat. At least use a reaction table like B/X where monsters could be helpful.
Also you need to use the print book and the side pdf to generate these encounters. You roll a % dice to see what type of encounter happens (or if it happens as there is often a 25% chance of no encounter) then a second % roll looking up what the monster is.
The monsters also seem to be chosen for their CR rating rather than their theme. They are a cornucopia, and the wandering monsters, mines, average, table we have Ankeg, Doppelganers, cocatrice, ettercap, hell hound, howler, Giant Mantis, Ogre, Rust Monster, Giant Scorpion, Shadow, Spriggan, Giant Wasp, Dire Wolf, Giant Stag Bettle, Fungal Crawler, Crab Swarm, Dark Stalker, Mimic, Minotaur, Giant Vulture, Dire Wolverine, and the four types of medium elementals. (and that is only 66% of the table!)
This is a particular shame as the adventure talks about the taint of the titular Dark Obelisk. There was every chance to grab a Dark Obelisk template, do up 2 dozen or so tainted monsters, and build a really good solid theme (just like Barrowmaze does with undead).
I can see how this is meant to work. Large dungeon crawl. Players fighting their way in and out. Slowly uncovering what is happening. But it's too big. The book is too big. The monster range for combats is too big. The headers take up too much space when the maps communicate the area so much better. Barrowmaze can be flipped through during play, this cannot. I like the base idea here but the execution needs a lot more work.