Not sure about the setup you are showing.
I assume you want to measure material properties of some sample at the bottom.
If these effects are too small in a normal waveguide run (travelling wave), in general the sample is placed in a cavity. The idea is that the Q of the cavity essentially amplifies the effect of the sample in a normal waveguide by the same factor Q. (Think of the wave bounding around approximately Q times, experiencing therefore Q times the effect of a normal terminated waveguide)
The cavity needs to be closed at both ends. The bottom seems shorted, however the top needs to be closed by either an iris or a post.
The length of the cavity is definitely of importance on the resonance. If you are using say WR90 waveguid, for this frequency range, it will support te TE10 mode. The numbers being the number of half sines along each dimension. For a cavity another dimension is added (the length). Also eg. a TE101 mode. Just use one of the online calculators. The length should be chosen so that the resonance is within your range.
For determining material properties, with a transmission measurement it is in general much easier to determine these, than from a reflection measurement.
Finally if you are worried about the gunnplexer experiencing full reflection, for hobby purpose, just glue a piece of blackened (by fire) wood on the side in the waveguide between the gunnplexer and the cavity, acting as a small attenuator.
BTW, lookup "microwave wavemeter" for frequency measurement.