Is it feasible to analyze ~10GHz structure using ~1GHz structure (size scaling idea)

Can we analyze passive microstrip / waveguide structures using this approach? Idea is to take 10GHz model, scale it to 10x bigger size to be as if it was 1GHz. Can we get some useful information from such 1GHz data Or diffirence will be too much?

In general, yes. Frequency scaling is useful in antenna design and elsewhere.
Be aware, however, that material permittivity, permeability and skin depth is a function of frequency, for 1:10 the change can be significant.

The most realistic scaling would be to scale the layout dimensions and the substrate thickness. This will maintain the effective dielectric constant, because the relative field distribution in substrate/air is the same. Also, radiation should be the same in that scaled model.

If you keep the substrate thickness, you need to keep the line width (line impedance) and scale only the line length. Coupling between line segments might change, so the scaled model is of limited value. I would prefer EM simulation of the original dimensions.

yep, back in the day people would design 94 GHz waveguide circuits with 1 GHz waveguide. The diodes were the size of a small coke can, lol.

There are some odd scaling rules as I recall.