[sekun]

I don’t understand what metrics you use for “more scalable” and “more maintainable”. If your application’s needs is sufficiently fulfilled by RLS, you don’t have to reinvent the wheel in the application level. Less code there to maintain is good.

Plus, I don’t see how “just by reading application code you can tell the expected behavior” doesn’t apply to RLS. Policies are written in a consistent format. USING for visibility, WITH CHECK for altering. I only have to keep an eye for these, and I’ll already get a good summary on what it does, no?