Practice Project — A sandbox to explore modern full‑stack patterns.

/project
├── backend/   # NestJS API: MongoDB/Mongoose, DTOs, JWT auth (access & refresh), token revocation workflow, file uploads
├── frontend/  # Next.js app: Axios access/refresh handling, React Query caching & optimistic updates, Tailwind, Yup validation, route guards, AuthContext
└── README.md  # You are here—it links to sub‑project READMEs

git clone <repo-url>
cd project
cp .env.example backend/.env
# edit backend/.env with MONGO_URI, JWT_ACCESS_SECRET, JWT_REFRESH_SECRET, etc.

# Backend
cd backend && npm install
# Frontend
cd ../frontend && npm install

# Start backend (NestJS)
cd backend && npm run start:dev
# Start frontend (Next.js)
cd frontend && npm run dev

• Backend → http://localhost:3000
• Frontend → http://localhost:3004

• Backend: Schema‑driven MongoDB with Mongoose, DTO validation, modular services (Bcrypt, Token), JWT access & refresh strategies with refresh‑revoke workflow, cloud file storage (Cloudinary), API rate‑limiting (Redis), and MongoDB query caching for high‑traffic endpoints (Redis)
• Frontend: Secure token management with Axios interceptors, automatic retry on 401, React Query hooks with cache invalidation and optimistic updates + rollbacks, Tailwind styling, Yup form validation, Next.js middleware and route protection, global AuthContext.

The backend API was evaluated under simulated production load using k6.

Configuration:

• Environment: Hosted backend (Heroku) with production configuration
• Load Profile: 50 virtual users (VUs) over 1 minute sustained traffic
• Endpoints Tested: POST /auth/login (initial auth), GET /list/items (authenticated resource)
• Authentication: JWT (access token valid through full test duration)
• Rate Limiting: Temporarily increased for load evaluation

Results:

• Total Requests: 1,526 successful API calls
• Success Rate: 99.6% (3040/3051 checks succeeded)
• Throughput: ~21 requests/second sustained
• Latency:
  • Average: 152 ms
  • p90: 161 ms
  • p95: 173 ms
  • Max: 965 ms (isolated outlier)

Summary:

API sustained 50 concurrent virtual users with 99.6% success rate and p95 latency under 175 ms during sustained load test.

Configuration:

• Environment: Hosted backend (Heroku) with production configuration
• Workflow Simulated: Login → Get Items → Add Item → Delete Item
• Load Profile: 50 virtual users (VUs) executing full workflow for 90 seconds

Results:

• Total API Calls: 3,295 across all CRUD actions
• Success Rate: 100% (all login, GET, POST, DELETE checks succeeded)
• Throughput: ~36 requests/second sustained
• Latency:
  • Average: 298 ms
  • p90: 786 ms
  • p95: 818 ms
  • Max: ~1.02 s (isolated slow request during writes)

Summary:

API sustained 50 concurrent authenticated workflows (login, create, fetch, delete) with 100% success rate and p95 latency <820 ms during production‑hosted workflow test.

Configuration:

• Environment: Hosted backend (Heroku) with production configuration
• Load Profile: 20 virtual users (VUs) sustained over 30 minutes
• Endpoint Tested: GET /list/items (authenticated resource)
• Authentication: JWT (access token valid through full test duration)

Results:

• Total API Calls: ~22.5k requests
• Success Rate: 99.98% (only 4 transient connection resets)
• Latency:
  • Average: 162 ms
  • p95: 232 ms
  • Max: 777 ms (isolated outlier)

Summary:

API sustained 20 concurrent virtual users for 30 minutes (~22.5k requests) with 99.98% success rate and p95 latency <235 ms, showing no degradation in performance or stability.

Both tests are implemented in k6 and included in the repo.

Prerequisites:

• Install k6
  • Mac (Homebrew): brew install k6
  • Windows (winget): winget install k6
  • Windows (Chocolatey): choco install k6

Test Script Locations:

• backend/src/test/load_test.js # Sustained load test
• backend/src/test/workflow_test.js # Full CRUD workflow test
• backend/src/test/soak_test.js # ~30min soak test

Run a Test: From the backend root:

k6 run src/test/load_test.js
k6 run src/test/workflow_test.js
k6 run src/test/soak_test.js