@Lerner: I’m still a bit confused. What is 3500 K? If the gas is 3500K I don’t see a problem holding the electrodes as 1100 K or less with some creative thermal engineering. If the metal is 3500 K, I don’t understand the logic. Your calculation of the heat transfer from what I presume is the gas is likely inaccurate. The results appears to assume that you can remove heat at 1.5 kW/cm^2. All heat flow is driven by temperature differences in the system and the convection of the gas if the gas is the hottest element. Radiation is important for a time but convection could also play a role as the gas is extremely turbulent and should be directed at the chamber and the inner anode diameter. The decay of the electron population is also important. Electrons will carry away and radiate away heat.
My piece of advice is design a larger vacuum spool so you can address the shock problem with CF gasket leading to vacuum issues if you still believe it a problem. This will allow for a large gas buffer and throw away volume. If you are interest in dealing with gas based nucleation, I would suggest looked at our semiconductor friends. They have these problems and deal with them. Probably a good start.