They are very robust, a knock wouldn't change the filament resistance. More likely, because the resistance is very low anyway, the probe contact or even the meter cable resistance was also playing a part in the measurement.
There is actually very little inside a Magnetron, a shaped cavity, a heating filament and a strong magnet so the chances of something breaking are slim. They are simple devices, the heater is also the negative electrode and it emits electrons. The cavity is the positive electrode that attracts the electrons toward it. The magnet skews the electron flow so it passes over the entrance to the cavity and sets up an oscillation. It's the same principle as blowing over the neck of a bottle making a tone except with electrons instead of wind. The size of the cavity decides the frequency just like the size of the bottle would.
You should be able to measure resistance across the heating filament, it is only a wire, much like the filament in an incandescent light bulb. You should not be able to measure resistance to the cavity (anode) becuase there is a vacuum gap between it and the filament.
Brian.