My approach would be to combine pretty much what everyone else has said, and add annealing to the process:
- More walls, higher temp, bricklayers (the usual wank when you're more into 3d printing tech than design). Test if watertight and proceed to #2 when it obviously isn't
- Can I just buy the container?
- No? Then annealing could be a solution. Take the print from #1 and put it in a baking tray. Fill out every bit of air with sand to support the print. Bake the print at just above the glass temp for the filament for a enough time for the sand to get up to temp. Turn off oven and let the entire thing cool down slowly. Test the dimensions and watertightness.
- Realize that I didn't do a good job of supporting the structure and I spent too much material printing extra walls. Others have suggested coating the print, which will add material in a somewhat unpredictable manner. Is it acceptable in view of my tolerances, can I alter the design to make it acceptable? Print with minimal material and coat with epoxy.
- Get tired of the project and start comfort printing cute stuff for people you like...