image

Single-Database vs. Multi-Database Multi-Tenancy in Laravel

17 Nov, 2026 Abdelrahman Etry

When building a B2B SaaS platform in Laravel—whether managing multi-branch car showrooms, clinics, or enterprise portals—your earliest architectural decision is how to isolate tenant data. Both single-database (shared schema with tenant_id) and multi-database (isolated database per tenant) approaches work cleanly at 10 tenants. At 500 tenants, the wrong choice creates severe operational friction.


Approach 1: Single-Database (tenant_id Scoping)

Every tenant shares the same database and tables. Isolation is enforced at the application layer via Laravel Global Scopes and automatic model observers.

PHP Sample Code:

// app/Models/Scopes/TenantScope.php
class TenantScope implements Scope{
    public function apply(Builder $builder, Model $model): void{
        if ($tenantId = app(TenantManager::class)->id()){
            $builder->where($model->getTable() . '.tenant_id', $tenantId);
        }
    }
}


Where it shines:

  • Instant onboarding: Creating a tenant is a single INSERT into tenants.

  • Fast deployments: Schema migrations run once on a single database in seconds.

  • Cross-tenant analytics: Super-admin reporting across all tenants requires standard SQL queries.


The hidden traps:

  • Data leak risk: Raw queries (DB::table(...)), forgotten withoutGlobalScopes(), or misconfigured queue jobs can expose data across tenants.

  • Noisy neighbor problem: One tenant running a massive export query spikes CPU/IOPS for everyone else.

  • Composite indexing overhead: Every single index must lead with tenant_id (e.g., INDEX(tenant_id, status, created_at)), or MySQL falls back to scanning large index ranges.


Approach 2: Multi-Database (Database-per-Tenant)

Packages like stancl/tenancy dynamically switch Laravel’s default database connection (DB::setDefaultConnection(...)) based on the incoming subdomain or API token.


Where it shines:

  • Hard isolation: Accidental cross-tenant data leaks via SQL become structurally impossible.

  • Point-in-time recovery: Restoring a single tenant who accidentally deleted their catalog doesn't require rolling back other customers.

  • Compliance & enterprise readiness: Enterprise clients frequently demand physical data isolation and custom backup retention.


The hidden traps:

  • Migration bottlenecks: Running php artisan tenants:migrate across 2,000 databases can turn a 2-second deploy into a 15-minute pipeline bottleneck.

  • Connection pool exhaustion: PHP workers opening connections to hundreds of distinct databases can overwhelm MySQL max_connections unless you place ProxySQL or PgBouncer in front of the database.


Architectural Decision Rule

Start with Single-Database if you are targeting high-volume SMBs with uniform schemas and rapid iteration. Choose Multi-Database (or a hybrid model where enterprise tiers get dedicated databases) when compliance, per-tenant backups, or heavy per-tenant write loads outweigh deployment simplicity.