◎ CloudPulse

Intelligent Serverless Website Monitoring & Alerting Platform

CloudPulse is a production-grade, fully serverless uptime monitoring system built on AWS. Track website availability, measure latency, get smart alerts, and view analytics — all with zero servers to manage.

---

Architecture

EventBridge (5min schedule)
        │
        ▼
Lambda: Health Checker ──────────────── CloudWatch Metrics & Logs
        │
        ├──▶ DynamoDB: cloudpulse-sites      (site registry)
        ├──▶ DynamoDB: cloudpulse-metrics    (time-series, 30-day TTL)
        └──▶ DynamoDB: cloudpulse-incidents  (incident history, 90-day TTL)
             │
             └──▶ SNS ──▶ Email
                      └──▶ Slack Webhook

API Gateway ──▶ Lambda: API Handler ──▶ DynamoDB
                │
React Dashboard ◀── Frontend (S3 + CloudFront)

---

Features

Feature	Details
Multi-site monitoring	Track unlimited URLs with per-site configuration
Intelligent health checks	HTTP status, latency, 2-retry logic, 5s timeout
Smart alerts	Email (SNS) + Slack, cooldown period prevents spam
Latency tracking	Time-series graphs, avg/p95 latency, 30-day history
Uptime calculation	Per-site % uptime with visual 30-day bar
Incident history	Tracks when sites went down and recovered
CloudWatch integration	Custom metrics, execution logs, built-in dashboard
Cost optimized	Serverless auto-scale, DynamoDB on-demand, TTL cleanup

---

Tech Stack

Backend (Serverless)

• AWS Lambda (Python 3.12)
• Amazon EventBridge (scheduling)
• Amazon DynamoDB (on-demand billing + TTL)
• Amazon SNS (email alerts)
• Amazon API Gateway
• Amazon CloudWatch (metrics + logs)

Frontend

• React 18 + Vite
• Tailwind CSS
• Recharts (latency graphs)
• Lucide icons

DevOps

• AWS SAM (infrastructure as code)
• GitHub Actions (CI/CD)

---

Project Structure

cloudpulse/
├── backend/
│   ├── lambda/
│   │   ├── health_checker/
│   │   │   └── handler.py
│   │   └── api_handler/
│   │       └── handler.py
│   └── tests/
│
├── frontend/
│   └── src/
│
├── docs/
├── .github/
├── template.yaml
└── README.md

---

Quick Start

Prerequisites

• AWS CLI configured (aws configure)
• AWS SAM CLI installed
• Python 3.12, Node.js 20+

---

1. Clone & install

git clone https://github.com/tejas-0-5/cloudpulse.git
cd cloudpulse

---

2. Deploy backend

# Build
sam build

# Deploy
sam deploy --guided
# → Stack name: cloudpulse
# → AWS Region: ap-south-1
# → AlertEmail: your@email.com

After deploy, SAM prints your API endpoint:

Outputs:
ApiUrl: https://your-api-id.execute-api.ap-south-1.amazonaws.com/Prod/

---

3. Deploy frontend

cd frontend

# Set API URL
echo "VITE_API_URL=https://your-api-id.execute-api.ap-south-1.amazonaws.com/Prod/" > .env

npm install
npm run build

aws s3 sync dist/ s3://YOUR-BUCKET-NAME/ --delete

---

4. Run locally

# Frontend dev
cd frontend && npm run dev

---

API Reference

Method	Path	Description
POST	/sites	Register a new site
GET	/sites	List all monitored sites
DELETE	/sites/{id}	Remove a site
GET	/metrics?site_id=X&hours=24	Time-series latency data
GET	/incidents?site_id=X	Incident history
GET	/dashboard	Summary stats for all sites

---

Configuration

Variable	Default	Description
LATENCY_THRESHOLD_MS	2000	Alert threshold in milliseconds
SLACK_WEBHOOK_URL	""	Slack webhook URL
SNS_TOPIC_ARN	auto	Set by SAM

---

Running Tests

pip install pytest moto[dynamodb,sns] boto3 urllib3
pytest backend/tests/ -v

---

CI/CD (GitHub Actions)

Set these secrets:

• AWS_ACCESS_KEY_ID
• AWS_SECRET_ACCESS_KEY
• ALERT_EMAIL
• S3_BUCKET_NAME
• CLOUDFRONT_DISTRIBUTION_ID

---

Cost Estimate (AWS)

~$0.40/month for 10 sites (low usage)

---

License

MIT

---

👨‍💻 Author

Tejas Pathak https://github.com/tejas-0-5