Skip to main content

Intro

Objective​

To build a scalable, secure, and efficient multitenant SaaS database architecture that supports seamless onboarding and data isolation for multiple tenants and products.

Approach: Multitenant Architecture​

We have adopted a silo-shard concept for our multitenant SaaS platform. Each tenant is provisioned with a dedicated silo, which is created dynamically in Google Kubernetes Engine (GKE) using a custom Helm chart. Each silo manages all data for a particular tenant across all their subscribed products, ensuring strong data isolation and scalability.

All metadata and database configuration details are maintained in a separate Aventian metadata database (out of scope for this document).

Architecture Diagram​

Tech Stack Used​

  • PostgreSQL: Primary database engine for tenant silos

  • Docker: Containerization of database and application services

  • Google Kubernetes Engine (GKE): Orchestration and management of silos

  • Kubernetes (k8s): Underlying container orchestration

  • Helm: Custom charts for automated silo provisioning

  • Flyway: Database migration and version control

For more details, refer to the specific product and resource documentation in this section.