About This Architecture

Mynime SaaS platform uses a multi-cluster EKS architecture with separate nonprod and prod environments, each running 16+ microservices across namespaces via Nginx Ingress and NATS message bus. Traffic flows through Route 53, AWS WAF, and ALB to fe-mynime (Next.js 14) and api-gateway (NestJS), with backend services including authorization, claims, insurance, and property-management microservices backed by CNPG and RDS Aurora PostgreSQL. Karpenter auto-scales nodes, cert-manager and External Secrets Operator manage certificates and secrets, while Promtail feeds observability data to CloudWatch and VPC Flow Logs. The architecture consolidates 16 services into 4-5 starter services, integrates Supabase for auth and realtime, AWS Bedrock for AI, and LlamaParse for document processing. Fork this diagram on Diagrams.so to customize cluster topology, add additional services, or adapt the multi-tenant database strategy for your SaaS requirements.

People also ask

How do you design a multi-tenant SaaS platform on AWS EKS with separate nonprod and prod clusters, microservice consolidation, and auto-scaling?

Mynime's architecture uses dual EKS clusters (nonprod and prod) with Nginx Ingress, NATS broker for async messaging, and 16+ NestJS/FastAPI microservices. Karpenter auto-scales nodes, CNPG and RDS Aurora provide data isolation per service, and Supabase handles auth and realtime. This pattern enables cost-efficient multi-tenancy while consolidating services from 16 to 4-5 starter services.