The downside of reef building?

I was reviewing text for a website a few weeks back dealing with aspects of the history of life, and a couple of things struck me about biological reefs. First, a quick look at reef-builders through time (a useful overview can be found here if you'd prefer more flesh on the bones, or corallites or shells or whatever may be more appropriate):

The first multicellular reef-builders were the archaeocyathan sponges, who flourished briefly in the Cambrian but did not even make it to the end of the period. Corals, in the form of rugose and tabulate corals, spread in the Ordovician but took a while to make reefs. They were joined by stromatoporoid sponges (layered like stromatolites, spelled like stromatolites, but not stromatolites) and various microbes, with the Devonian being an apex of reef-building. The classic stromatoporoid-tabulate reefs of the Devonian went kaput in the End-Devonian extinction. Permian reefs were a conglomeration of just about everything that couldn't get out of the way: various algae, sponges, bryozoans, and other less obvious things. This assortment bought it at the end of the Permian. False starts with scleractinian corals in the first part of the Mesozoic gave way to the rudist bivalves in the Cretaceous. The rudist reefs went out with non-avian dinosaurs, marine reptiles, pterosaurs, ammonites, and so forth at the end-Cretaceous extinction. Finally we get to the big scleractinian coral reefs in the Cenozoic, with some sponge and oyster reefs and such for variety.

So far, a typical pattern: group of organisms branches into reef building, reefs spread and become ecologically complex, reefs flourish for a while, mass extinction wipes out reefs. Then after a hiatus reefs become fashionable again, with some other group laying the foundation for a new iteration, and the cycle continues. A couple of observations come to mind. First, reefs seem to be an obvious evolutionary path for immobile marine invertebrates. It may take some time, but some group always takes up the baton after another falters.

Then, the other shoe. What happens to the previous reef-builders? Seen any archaeocyaths lately? Any vacation packages advertising stromatoporoid reef visits for their island getaways? Run across any rudists while snorkeling? Could it be that once a group goes all-in on the reef habit, it's stuck with it?

Furthermore, reefs have a habit of getting smacked in mass extinction events. Does a reef inherit a narrowing range of environmental restrictions from its components as its complexity increases? Does it become vulnerable to unpredictable instability, such as some minor constituent going through a bad patch leading to collapse via a Rube Goldbergian-cascade of events? More broadly, does reef building amount to an evolutionary Faustian bargain, in which a group becomes dominant for a while by locking itself into a doomed arrangement? (Granted, we're all doomed in the final analysis, but some of us are more obviously doomed than others.) Or am I just playing the gloomy Minnesotan?