HTTP Backends (Java & Node)

Parallel backends in Spring Boot and Express that expose the same REST surface: CRUD, validation, JWT auth, pagination, rate limiting, background jobs, OpenAPI, k6 load tests—Dockerized with Postgres/Redis and basic observability.

JavaSpring BootNode.jsExpressPostgreSQLSupabaseRedisDockerNGINXk6JWT

View Repository →

The Goal

Build the same HTTP API twice—once in Spring Boot (Java) and once in Express (Node)—to compare ergonomics, performance, testing strategies, and deployment. Both backends share Postgres and Redis and are exercised with identical curl, k6, and integration tests.

Solution Overview

Common REST surface: /health, /todos, /bookmarks
Full validation, pagination, filtering, and sorting parity
JWT auth + refresh tokens; rate limiting via Redis
OpenAPI spec, k6 load tests, and Dockerized local stack

Architecture & Stack

Node.js Service

Express + Zod validation
pg client for Postgres
Redis cache + rate limiting
Jest integration tests

Java Service

Spring Boot (Web, Data JPA)
Hibernate → Postgres
JWT filter + Controller/Service/Repo
Maven + Spring Test

[Architecture Diagram Placeholder]

Java & Node services → NGINX → Postgres + Redis

Key Endpoints (Shared Contract)

GET /health – service + DB + cache status
GET /todos?limit=&offset= – pagination
POST /todos, PATCH /todos/:id, DELETE /todos/:id
/bookmarks – filtering, sorting, search
/auth/login, /auth/refresh

# Create
curl -X POST localhost:4000/todos \
  -H 'content-type: application/json' \
  -d '{"title":"learn http","done":false}'

Tests, Perf & Observability

Integration tests for success & error paths; OpenAPI contract checks
k6 load tests; compare p95 latency Java vs Node
Request ID logging, basic metrics (/metrics)

Deployment

Docker Compose for local dev (Postgres + Redis + both services). CI builds and runs tests; images can be pushed to a registry and deployed behind NGINX as an API gateway.