I doubt the tanks will be the root cause of any future payload losses now that they have identified them as a vulnerability. They probably have a vastly better understanding of the envelope of that system now.
SpaceX has basically said that the voids are fine as long as the oxygen does not freeze. If that is true then the "nitroglycerine" in your analogy would be the incredibly low temperatures combined with the fast loading times that they were using. (They were obviously loading helium that was cold enough to freeze oxygen... Yeah. Seriously.)
Now that SpaceX has the most economical launcher on the market I think their strategy is going to change from "launch often - fail quickly", to something like "test often - fail quickly on the testbench". Now that they have a rocket that can do work and make money, a rocket that was funded partly with tax money to carry people for NASA, their focus will naturally shift towards making it more reliable.