I saw a question asking if metals sublimate, with some fascinating answers, which made me wonder if sublimation was the main cause of the eventual, and inevitable failure of vacuum tubes. High school electronics teachers attributed failure to the slow oxidation of the filament coating, but tubes are highly evacuated, and then a “getter” is fired inside the new sealed tube which (by means that escape memory; been doing solid state for 50 years now) traps errant gas atoms. It would stand to reason that an evaporating filament is more likely to explain the slow degradation of performance during the service life of the tube.
Yes. Sublimation weakens and thins the filament in vacuum tubes (and lightbulbs). You see this metal condensing as a metallic sheen on the inside of the "cooler" glass at the tops of the "radio vacuum tube" and overall inside the light bulb. The inrush of current when power is applied heats the filament, but the narrowest portion segment has the highest differential resistance. If the heat is sufficient to melt the narrowing (over time) portion of the filament it fails. Generally, "old" lightbulbs burn out when they are turned on. If power is applied more slowly it can prolong their lifespan. In the 70's, there was a "quarter" (USA 25-cent piece) sized disc that could be placed in the lightbulb socket, below the lightbulb base (it had insulation around the periphery so as to prevent shorting to the shell of the lightbulb socket). This disc had "high" resistance that reduced as it self heated. This had the effect of slowly increasing the voltage to a lightbulb and was supposed to increase the lightbulb's lifespan.
Vibration of the filament contributes to failure, either internal or external. This too can break "weakened" filaments. There is also flexing of the filament during the heating phase, because of rapid thermal expansion, and vibrations are set-up through this process which dampen out slowly.
There may be magnetic flux interaction in a wound filament that contributes to background vibration of a wound filament, but this is beyond my understanding.
If you hold a magnet next to a filament style lightbulb that is producing light by having been powered on, you can cause the filament to gyrate "wildly" through the interaction of the magnetic fields in Alternating Current circuits. Early flickering incandescent filament lightbulbs had a magnet built into the bulb near the filament. Interaction with the Earth's magnetic field may produce background vibration.