“At what point did you start moving logic out of RLS into your backend? Was it complexity, debugging, or something else?”
The user is inquiring about the challenges of managing logic within Row Level Security (RLS) and considering moving it to the backend. They are curious if complexity or debugging issues prompted others to make similar changes.
I never did, once you start creating functions to handle permissions instead of trying to write raw queries directly inside of the RLS policy themselves it becomes a bit more manageable.