StyleBetter
Microservices-Based Fashion Platform
Microservices backend for fashion and style app. gRPC-based services with polyglot persistence and presigned uploads.
Impact — P99 latency under 300ms across services; 3x faster catalog and search after Redis and query optimization.
Problem
A fashion platform needed independently deployable services, faster catalog/search, and reliable media uploads without coupling every domain into one monolith.
Approach
Node.js/TypeScript microservices with gRPC between services, PostgreSQL + MongoDB persistence, Redis caching, RabbitMQ for async notifications, and S3/CloudFront for uploads.
What shipped
- Built multiple Node.js/TypeScript microservices; JWT auth and user management for 100K+ users
- gRPC service-to-service communication; reduced inter-service latency by ~40% vs REST
- PostgreSQL + MongoDB persistence; Redis caching—3x improvement on catalog and search response time
- RabbitMQ for async notifications; S3 + CloudFront presigned uploads with reduced upload failures
- Swagger docs and multi-environment deployments; faster releases and fewer production incidents
Stack
Related writing
- NestJS vs Express vs Hapi vs Koa: When to Pick Which
A practical comparison of popular Node.js frameworks for APIs and microservices—structure, ergonomics, and team fit.
- Redis for Cache, Locks, and Rate Limiting
Three production patterns that cut database load, prevent race conditions, and protect APIs from abuse.