Dang. Of course the answer to that isn’t completely straightforward.
The column is air tight. My first prints using non-dried filament were more porous than later prints using dried filament. I still have painted them with thick cyanoacrylate to seal them though. You’d be surprised how porous 3d prints can be though. I increased walls/ceilings/floors layers and went with honeycomb infill (gyroid is very open) to help with leaks as well. My last print was very tight but not perfect. Leaking a little ok but easy enough to make it totally tight.
About those beads…That’s another area that’s luckily not terrible in practice. Because the water sticks down pretty hard to the silica gel bead surface area, it tends to stay where it sticks - mostly. As more water lays down, additional water is held less tight. As more water collects at the reservoir entrance it gets more mobile. As it uses its new-found mobility to move down the column it gets to drier and drier beads until it sticks down again. It’s just in its new home it’s held tighter and is less likely to get up and move. This is the process that sets up that gradient you saw in the photo.
That process continues whether the pump is running or not. It’s just with the pump off there is no bulk flow to move the water through the column. It’s all random walk. But the process continues. All those beads with a range of indicator colors in the nice gradient pattern got dumped back into the desiccant container to be regenerated later. All these bright orange beads next to dark green. It was definitely easy to see who held water and who didn’t.
I looked back in after a couple of weeks and they were all an intermediate darkish green. The water had equilibrated to an average value over all the beads.
The path was short. It was just bead to adjacent bead for the most part with no gradient to hold moisture at bay. It shows what happens over time, though. If the column just sits, how bad that sitting is depends on how much water it’s holding. If it’s mostly dry there won’t be much effect. If it’s got a lot of water the dark band at the entrance will gradually blur out during long storage and if if reaches the end of the column, you’ll start having exit humidity start rising prematurely. I have little data to know how fast it happens but I’m still using that then new column in the photo. That’s the one I tested and saw a 2% RH reading still a couple of weeks ago after I’ve been using it off and on since that photo dated September 2.
There’s another user here who built a small version of his own design with a 200g reservoir. His is working great last I heard. He just needed less spools dried and I think he said he is getting 5 or 6 spools dry using his. But again, how long the desiccant will last depends on your ambient humidity.