Staging-first SQL
All readiness SQL, migrations, views, indexes, and grants should be reviewed and executed first in staging. Production promotion requires explicit approval and rollback awareness.
This note makes database readiness controls visible for donor review. It is documentation only. It does not execute SQL, alter Supabase schema, change RLS policies, modify credentials, or affect live production logic.
All readiness SQL, migrations, views, indexes, and grants should be reviewed and executed first in staging. Production promotion requires explicit approval and rollback awareness.
Every protected route should be tested against role, organization, programme, reviewer, and admin boundaries. Tests should include allowed, denied, and cross-tenant cases.
High-use routes, exports, dashboards, MRV views, queues, and audit lookups should have supporting indexes and explain-plan review before donor-scale traffic.
Connection pooling should be checked for serverless functions, export bursts, background jobs, and authenticated workspace traffic.
Point-in-time recovery, backup retention, restore drills, recovery ownership, and incident communication should be confirmed before production scale-up.
Operational monitoring should cover database health, auth/session errors, function errors, failed exports, queue backlogs, background jobs, and suspicious access patterns.
Staging tests should cover login, protected navigation, evidence read/write, export pack generation, MRV summary reads, and admin/reporting flows.
AI Advice, exports, satellite refresh, MRV checks, notifications, and queue processing need status, retry, ownership, and failure visibility.
The presence of this note does not certify the database as production-ready. It documents the control domains donors and technical reviewers should verify before production deployment or donor-scale use.