Per your comments about stresses and strains I suspect bigger = better.

No, as Steve already stated, just the opposite!

The mass of a cube of water of 1m x 1m x 1m is m = 1000kg. Makes a pressure of P = F / A = m x g / A = 1000 x 9.81 / 1 x N/m/m = 9810 N/m/m, or nearly 10000N/m/m.
4000m under sea level pressure increases to 4000 x 10000N/m/m = 40MN/m/m.

Now assume that your circuitry is potted into a cube of 5cm x 5cm x 5cm. Then each surface is exposed to a force of F = P x A = 40MN/m/m x 5cm x 5cm = 100000N which corresponds to the weight of a mass of 10000kg!

What does this force cause?
A volume shrinkage! This volume shrinkage can be calculated, when bulk modulus K or modulus of elasticity E is known.
Problems arises, when your application contains materials with DIFFERENT E, because then relative forces inside the cube develop. E.g. PCB made of FR4 will show different E than potting material. If now SMD devices are soldered on PCB and whole application is potted in epoxy or polyurethan, then E is rather dfferent, relative forces develop and potting material is going to tear the SMDs from PCB!

Let's call 'l' the side length of the cube, means the cube is assumed to have volume l x l x l. Then l undergoes a shrinkage according to:

dl/l = -dp / E

where dl/l is the shrinkage in each dimension, dp is the applied pressure,

then FR4 PCB material with E = 18 x 10**9N/m/m undergoes a shrinkage of dl/l = -40MN/m/m / (18 x 10**9 N/m/m) = -0.22%. So, the cube with 5cm side length becomes smaller by about 111µm.

Epoxid potting material is more elastic than FR4, because it's not armed by glass laminate. E is different from manufacturer to manufacturer. If we take E = 9 x 10**9 N/m/m, then we get dl/l = -0.44%, means potting material undergoes a shrinkage of 222µm. Due to the big hardness of potting material enormous forces develop on the SMDs on PCB!!

Keep in mind, that when using bigger cubes dl/l remains constant, so that the absolute shrinkage increases. Assume your cube has 50cm (only for anylzing purpose), then absolute shrinkage of both materials relative to each other will be 1.1mm!! From this you can see, that volume of your application should be as minimal as possible.

There's another problem. If you have a look at SMD packages of ceramic capacitors e.g. you will notice, that the body does not directly touch the PCB when soldering the SMDs on PCB. Only the solder pads do touch the PCB, but even here a thin solder tin film is between. Unfortunately, the potting material is not going underneath the body, so that the body is hanging 'in the air'. If now this enormous pressure of 40MN/m/m is applied to application, ceramic SMD capacitors will break, in such a way, that body is touching PCB.
There are some ways to make the potting with vacuum and so on. But even with this technique it's often observed, that the space between SMD body and PCB is NOT filled with potting material...

By the way, potting electronics into hard epoxid material isn't trivial, either. Why?
Because even when hardening a certain relevant volume shrinkage of some % will be observed! So, your application can be damaged already after the potting and hardening procedure, long before being applied to high pressure.

I would highly recommend to put the circuit into a V2A cylinder, so that the PCB is not exposed to this enormous pressure.

Kai