Tenant Hosting - MQUAD Platform

MQUAD supports multiple tenant data isolation strategies to meet different compliance and scalability requirements.

Isolation Strategies

Strategy	Isolation	Cost	Complexity	Use Case
Shared Database	Row-level	Low	Low	Small tenants, SaaS
Schema-per-Tenant	Schema-level	Medium	Medium	Medium tenants
Database-per-Tenant	Database-level	High	High	Enterprise, compliance

Shared Database (Default)

All tenant data resides in the same database, isolated by tenant_id column:

# Every query filters by tenant_id
query = select(Survey).where(Survey.tenant_id == request.state.tenant_id)

Pros: Simple, low overhead, easy to manage.
Cons: No hard isolation boundary.

Database-per-Tenant

For maximum isolation, each tenant gets dedicated databases:

async def _create_tenant_databases_for_services(self, tenant_id: str):
    services = ["survey_db", "data_db", "workspace_db", ...]
    for base_db in services:
        tenant_db = f"{base_db}_tenant_{tenant_id.replace('-', '_')}"
        # CREATE DATABASE tenant_db
        # Run migrations on tenant_db
        # Register in tenant_metadata

Naming Convention

{base_db}_tenant_{tenant_id_underscored}

Example: survey_db_tenant_acme_org

Tenant Creation Flow

Create tenant record in appdb.tenants
Seed tenant roles (Admin, Editor, Viewer)
Create tenant databases across all services
Run Alembic migrations on each tenant database
Create admin user for tenant

Configuration

Tenant isolation mode is configured via environment variables:

TENANCY_ENABLED=true
TENANT_DB_ISOLATION=database  # shared | database

Frontend Routing

// Dashboard layout routes by tenant status
if (isDormant && isAdmin) -> /dormant-admin
if (isDormant && !isAdmin) -> /tenant-offline
if (!is_active) -> /non-active