The 999 wall count approach essentially eliminates all infill and replaces it with walls. It is quite a common work around to get more consistent 100% infill parts. However, even with increased line widths, it is not able to eliminate porosity.
As for coatings, I occasionally paint prints using Emaille color or use acryllic spray to apply a UV protective coating. These approaches would probably not be suitable for your application though.
Maybe a dip coating using epoxy could give a better performance, but that is quite actually quite difficult as coating thickness will vary even with multiple layers.
Using PETG may improve some aspects like longetivity when used outsides, but PETG has its own pitfalls. A TPU will probably improve some aspects (like interlayer bonding), but will probably allow too much movement under internal pressure. It is actually quite difficult to beat PLA in z-impact. Besides TPU, only PA6 is stated with a higher z-impact. By a mere 12%.
So there just are some basic limitations to FDM which can become quite extensive to overcome.
For workable prototypes, short term applications, etc, it may be realizable using PLA/PETG with TPU thrown into the mix with many thick lines. But I’d not consider FDM for long term use under internal pressure. And it is quite important not to expose such parts to freezing conditions. Their porosity means that they’ll have water insides its solid surfaces after a few weeks. You do not want that to freeze…
Alternatives are of course prototype development using FDM followed by classical milling/turning or, if “the way” is the goal, consider printing moulds for fibre reinforced epoxy.
& 