I put a transzorb across the VCC for the each group of IXDD414s. A series diode in series, at least with each group of IXDD414s +VCC line is not a bad idea at all - or if you're really paranoid, a diode in series with each IXDD414's supply.
I'm guessing you have transzorbs across the drain busses as well. You have an overload shutdown that kills the main power supply along with the pulse train.
There is nothing at all wrong with "belts and suspenders"
One thing that has really come far over the past few years is protection. The old days of the catatrophic failures are behind us. Think about it - there would have to be a lot of things failed, and all at the same time, to get any sort of cascading situation going - and even then, you've taken steps (as outlined above and in posts by Frank and others) to stop it. If something shorts out in the RF deck. The overload shutdown immediately kills the DC to the RF amplifier before the real current gets going. The instant the current rises faster than the voltage, the overload detect stops everything cold.
Those 18V transzorbs are amazing! Bob K1KBW, using analog drive, had his driver putting out some 80 watts or so into an 8 MOSFET transmitter. After a little while of this, the transzorbs finally gave up the ghost and shorted dead, but the gates were unscathed. I've never had anything get through them under any circumstances..... but there's always a first time...
I use seperate power supplies for each module's IXDD414s in my designs, with 18V transzorbs on each power bus.
Never had a failure of any kind whatsoever the RF amplifier, modulator or driver despite my best efforts to transmit at full power into no load (many times), or on the wrong band or into a dead short or with no drive.... The only thing I've had to replace were 2 50 ampere bridge rectifiers in the power supply... and this was due to an open step-start resistor. I put in a "real" resistor, and no more rectifier problems!!