Skip to content

Topic

Postgres Row-Level Security

PostgreSQL feature that restricts which rows a query can read or modify, using per-role policies often used for multi-tenant data isolation.

Current stories

build1 publisherOne report

Supabase ends automatic role grants on new tables in existing projects from October 30

Supabase stops auto-granting anon, authenticated and service_role on new tables in existing projects on October 30, an auditor writing on dev.to reports. Tables created after that need scoped grants in their migrations, since the common GRANT ALL TO public shortcut lets anonymous callers back in.

Publishers:dev.to

Reality

Evidence50
Adoption
Insufficient
Hype gap+10
Incentives55
Confidence55
build1 publisherOne report

Missing Column GRANT, Not RLS, Blocked an Admin Balance Top-Up in Postgres

Postgres blocked a coffee shop's admin top-up at the column GRANT check before RLS ran, and a pg_proc scan found four functions built the same way. The error text does not say which gate refused, so when a correct policy still denies a write, check the grant first.

Publishers:dev.to

Reality

Evidence50
Adoption
Insufficient
Hype gap0
Incentives
Insufficient
Confidence55
build1 publisherOne report

A Postgres role with no grants reads the FORCE flag on tables it cannot select from

PostgreSQL 18.3 handed a serving role with no grant at all on three tenant tables their exact owner and FORCE flags, in a dev.to author's test of three roles. The result gives an RLS setup check a third answer, UNDETERMINED, in cases where it would otherwise return a false nothing to report.

Publishers:dev.to

Reality

Evidence64
Adoption
Insufficient
Hype gap−10
Incentives25
Confidence60
build1 publisherOne report

A missing sequence GRANT broke production behind 21 working RLS policies

A solo-built WhatsApp CRM enforces its tenant boundary with row-level security on 21 tables, and the privilege that actually stopped a production insert sat on a table's auto-increment sequence, granted to one database role and not the other.

Publishers:dev.to

Reality

Evidence45
Adoption20
Hype gap−10
Incentives65
Confidence50