About This Architecture

Humalike's production API backend on AWS combines CloudFront edge distribution, WAF protection, and an internal ALB routing traffic across 12+ ECS Fargate microservices in a VPC Lattice service mesh. API requests flow through Route 53 DNS, CloudFront caching layer, and WAF Web ACL before reaching the ALB, which path-routes to specialized services like svc-python, svc-credits, svc-billing, and svc-ws-gateway. Async workloads are decoupled via EventBridge and SQS queues with dead-letter queues, while Aurora PostgreSQL Serverless v2 provides read/write database separation and Secrets Manager secures credentials. This architecture demonstrates high-availability patterns across two AZs with NAT gateways, VPC endpoints for private SaaS integrations, and observability via PostHog, Better Stack, and Doppler. Fork and customize this diagram on Diagrams.so to adapt the service mesh, add additional microservices, or modify the edge caching strategy for your own multi-tenant API platform.

People also ask

How do you architect a production AWS API backend with multiple microservices, edge caching, WAF protection, and async messaging?

This diagram shows a complete production setup: CloudFront and WAF protect the edge, an internal ALB routes requests via path-based rules to 12+ ECS Fargate services in a VPC Lattice mesh, Aurora PostgreSQL Serverless v2 handles data with read/write separation, and EventBridge plus SQS decouple async workloads. Multi-AZ deployment with NAT gateways and VPC endpoints ensures high availability and s