Skip to content
Docs home

SECURITY · UPDATED AUGUST 2026

Tenant Isolation

Row-level security as the enforcement point: every query runs under the tenant's role, and the application never assumes authorization the database did not verify.

Where isolation is enforced

Tenant isolation in QUE is enforced at the database through row-level security. Every query the application runs executes under the tenant's JWT role, and the application service layer never assumes authorization that the database did not verify. This matters because it removes the application's ability to accidentally trust itself: even a buggy server function cannot read across tenant lines if the database policy is correct.

The verification procedure

Isolation is verified through a two-tenant test: create two tenants, use separate identities, attempt permitted reads and forbidden cross-tenant reads from each, and check that the forbidden attempts receive errors rather than data. The result is recorded without exposing real customer data. The test suite runs this as part of the release gate, so an isolation regression blocks release rather than surfacing in production.

Isolation is a property, not a promise

Documentation describes the intended boundary, the test suite, the role policy, and the review procedure. It does not claim perfect isolation or an absolute guarantee — no real system can. What it claims is that the boundary is database-enforced, continuously tested, and reviewable.

STILL READING

Something here still feels worth checking together.

This documentation describes the product exactly as it is implemented. If a behavior you expected is not here — or is here and does not match your deployment — tell the team. Implementation questions start on the contact page.