Distributed Load Balancer

A custom HTTP load balancer built from scratch in Node.js with zero dependencies. Features round-robin routing, active health checks, real-time metrics, and a traffic simulator. Deployable both locally via Docker and to the cloud via Render.

---

Project Layout

load-balancer-project/
├── backend/
│   ├── server.js              # HTTP backend server
│   ├── package.json
│   └── Dockerfile
├── loadbalancer/
│   ├── load_balancer_health.js  # Round-robin LB with health checks + metrics API
│   ├── package.json
│   └── Dockerfile
├── simulator/
│   ├── traffic_simulator.js   # Concurrent traffic generator
│   ├── package.json
│   └── Dockerfile
├── dashboard/
│   ├── dashboard.html         # Real-time metrics UI
│   ├── serve.js               # Static file server
│   ├── package.json
│   └── Dockerfile
├── docker-compose.yml
├── .dockerignore
└── README.md

---

Architecture

Internet / Traffic Simulator
          │
          ▼
  Load Balancer :8080          (Round Robin + Health Checks)
  Metrics API   :8080/metrics  (JSON snapshot)
    ┌─────┼─────┐
    ▼     ▼     ▼
 :5001  :5002  :5003
server1 server2 server3

Dashboard :3000 → polls /metrics → live UI

---

Port Map

Service	Local Port	Render URL
server-1	5001	https://lb-backend-server-11.onrender.com
server-2	5002	https://lb-backend-server-2.onrender.com
server-3	5003	https://lb-backend-server-3.onrender.com
load-balancer	8080	https://lb-load-balancer.onrender.com
metrics API	8080/metrics	https://lb-load-balancer.onrender.com/metrics
dashboard	3000	https://lb-dashboard.onrender.com

---

Environment Variable Reference

Backend servers

Variable	Default	Description
SERVER_ID	server-1	Identifier returned in responses
PORT	5001	Port to listen on

Load balancer

Variable	Default	Description
BACKEND_1	localhost:5001	Full URL of backend server 1
BACKEND_2	localhost:5002	Full URL of backend server 2
BACKEND_3	localhost:5003	Full URL of backend server 3
PORT	8080	Port for LB + metrics API

Traffic simulator

Variable	Default	Description
TARGET_URL	http://localhost:8080	Load balancer to target
TOTAL_REQUESTS	1000	Total requests to send
CONCURRENCY	50	Max in-flight at once

---

Option A — Run Locally with Docker

Prerequisites

• Docker Desktop installed and running

1. Build all images

docker-compose build

2. Start the full stack

docker-compose up

Or detached (background):

docker-compose up -d

3. Verify everything is healthy

docker-compose ps

Expected:

NAME             STATUS          PORTS
server-1         Up (healthy)    0.0.0.0:5001->5001/tcp
server-2         Up (healthy)    0.0.0.0:5002->5002/tcp
server-3         Up (healthy)    0.0.0.0:5003->5003/tcp
load-balancer    Up (healthy)    0.0.0.0:8080->8080/tcp
dashboard        Up              0.0.0.0:3000->3000/tcp

4. Test round-robin

# 9 requests — should rotate across all 3 servers
for i in {1..9}; do curl -s http://localhost:8080 | grep server_id; done

5. Open the dashboard

http://localhost:3000

6. Run the traffic simulator

# Default: 1000 requests, 50 concurrent
docker-compose --profile simulator run --rm simulator

# Heavy load: 5000 requests, 100 concurrent
docker-compose --profile simulator run \
  -e TOTAL_REQUESTS=5000 \
  -e CONCURRENCY=100 \
  --rm simulator

7. Simulate a server failure

docker stop server-2
# Watch LB logs — server-2 marked DOWN, traffic redistributes with zero failures
docker start server-2
# server-2 automatically rejoins rotation within 3 seconds

Useful Docker commands

docker-compose logs -f load-balancer   # tail LB logs
docker-compose logs -f server-1        # tail a backend
docker-compose build load-balancer     # rebuild one image
docker-compose down                    # stop everything
docker-compose down --rmi all          # full clean

---

Option B — Deploy to Render (Cloud)

Prerequisites

• GitHub account with this repo pushed
• Render account at https://render.com (free tier)

Deploy order

1. Deploy all three backend servers first
2. Deploy the load balancer with backend URLs as env vars
3. Deploy the dashboard

Backend servers (repeat 3 times)

Setting	server-1	server-2	server-3
Name	lb-backend-server-1	lb-backend-server-2	lb-backend-server-3
Root Directory	backend	backend	backend
Start Command	node server.js server-1 $PORT	node server.js server-2 $PORT	node server.js server-3 $PORT
Env: SERVER_ID	server-1	server-2	server-3

Load balancer

Setting	Value
Name	lb-load-balancer
Root Directory	loadbalancer
Start Command	node load_balancer_health.js
Env: BACKEND_1	https://lb-backend-server-1.onrender.com
Env: BACKEND_2	https://lb-backend-server-2.onrender.com
Env: BACKEND_3	https://lb-backend-server-3.onrender.com

Dashboard

Setting	Value
Name	lb-dashboard
Root Directory	dashboard
Start Command	node serve.js

Test the live deployment

# Single request through the load balancer
curl https://lb-load-balancer.onrender.com

# Round robin check — 6 requests
for ($i=1; $i -le 6; $i++) {
    curl https://lb-load-balancer.onrender.com
    Start-Sleep -Milliseconds 300
}

# Live metrics JSON
curl https://lb-load-balancer.onrender.com/metrics

Run simulator against live Render deployment (from local machine)

cd simulator
$env:TARGET_URL = "https://lb-load-balancer.onrender.com"
node traffic_simulator.js 500 20

Test failover on Render

1. Go to Render dashboard → click a backend service → click Suspend
2. Watch the load balancer logs — server marked DOWN within 3-5 seconds
3. All traffic seamlessly redirects to remaining backends, zero failures
4. Click Resume — server rejoins rotation automatically

---

Key Features

• Zero dependencies — built entirely on Node.js built-in modules (http, https, url)
• Round Robin routing — even traffic distribution across all live backends
• Active health checks — pings every backend every 3 seconds, removes dead servers instantly
• Automatic failover — failed requests retry on next live backend transparently
• Auto-restart — Docker restart: unless-stopped and Render auto-restart keep services running
• Real-time dashboard — live bar charts, latency tracking, UP/DOWN badges, event log
• Cloud-ready — HTTPS support, Render $PORT env var, env-configurable backend URLs