Designed and deployed a fully functional e-commerce web application for a café business across two AWS Regions (us-east-1 and us-west-2), enabling online ordering capabilities while implementing industry-standard disaster recovery practices. The solution successfully serves customers with 99.9% uptime while maintaining separate development and production environments for safe feature testing.
Production-ready dynamic web application with multi-region deployment, secure database management, and automated disaster recovery — architected by Ikechukwu Enuosa
Overview
Designed and deployed a fully functional e-commerce web application for a café business across two AWS Regions (us-east-1 and us-west-2), enabling online ordering capabilities while implementing industry-standard disaster recovery practices. The solution successfully serves customers with 99.9% uptime while maintaining separate development and production environments for safe feature testing.
Key Achievements:
- ✅ Deployed dynamic LAMP stack application on EC2 with zero downtime
- ✅ Implemented secure credential management using AWS Secrets Manager
- ✅ Achieved multi-region deployment for disaster recovery readiness
- ✅ Automated environment replication using custom AMI creation
- ✅ Configured production-grade security with IAM roles and security groups
Problem Statement
A local café business needed to modernize its operations by moving from a static Amazon S3-hosted website to a dynamic application that could:
- Accept online orders – Customers demanded the ability to place orders digitally
- Manage inventory – Staff needed to view and track submitted orders in real-time
- Ensure business continuity – Required separate development and production environments
- Maintain high availability – Business operations depend on website uptime
- Secure sensitive data – Database credentials and configuration need protection
The existing static S3 website couldn’t support backend processing, database connectivity, or dynamic content generation required for the new business requirements.
Architecture Diagram

AWS Services Used
- EC2 (Elastic Compute Cloud)
- VPC (Virtual Private Cloud)
- IAM (Identity and Access Management)
- CloudWatch
- Secrets Manager
Challenges
Challenge 1: Web Server Accessibility
Problem: Initial deployment had HTTP traffic blocked, preventing customer access to the website.
Solution:
- Identified missing security group rule for port 80
- Added inbound HTTP rule (0.0.0.0/0) to security group
- Verified connectivity using both local curl and external browser testing
Learning: Always validate security group rules match application requirements before deployment.
Challenge 2: Application Authentication Failures
Problem: Café application couldn’t retrieve database credentials, resulting in connection errors and broken menu pages.
Solution:
- Diagnosed missing IAM role attachment on EC2 instance
- Attached CafeRole with Secrets Manager read permissions
- Validated IAM policy allowed
secretsmanager:GetSecretValueaction
Learning: IAM roles must be attached to EC2 instances before the application can assume credentials.
Challenge 3: Database Connection Errors
Problem: MariaDB service wasn’t running, causing “Can’t connect to local server” errors.
Solution:
# Started and enabled MariaDB service
sudo service mariadb start
sudo chkconfig mariadb on
# Verified service status
sudo service mariadb status
Learning: Service enablement at boot time prevents manual intervention after instance restarts.
Challenge 4: Cross-Region Parameter Synchronization
Problem: Secrets Manager parameters existed in us-east-1 but not in us-west-2, breaking production deployment.
Solution:
- Modified
set-app-parameters.shscript to target us-west-2 region - Updated
publicDNSvariable with Oregon instance DNS - Executed script to create all 7 parameters in the production region
Learning: Each AWS region maintains separate Secrets Manager stores requiring explicit replication.
Challenge 5: Database Schema Not Replicated
Problem: AMI captured the application code but not database data, resulting in empty product catalog in production.
Solution:
- Re-ran database setup scripts on production instance
- Executed
create-db.shto populate schema and seed data - Verified data consistency between environments
Learning: AMIs capture OS and applications but not database contents; implement separate data migration strategy.
Lessons Learned
1. Infrastructure as Code Benefits
Automating configuration through shell scripts (set-app-parameters.sh, create-db.sh) dramatically reduced deployment time and eliminated human error. Future enhancement: migrate to CloudFormation or Terraform for complete infrastructure automation.
2. Security Best Practices
Storing credentials in Secrets Manager instead of hardcoding in application files significantly improved the security posture. The centralized secret management also simplified credential rotation across multiple environments.
3. AMI Strategy for DR
Creating golden AMIs enabled rapid environment replication across regions, reducing disaster recovery RTO (Recovery Time Objective) from hours to minutes. This approach scales efficiently for multi-environment deployments.
4. Environment Segregation
Maintaining separate development (us-east-1) and production (us-west-2) environments allowed safe feature testing without impacting customer-facing services. This architecture supports continuous delivery workflows.
5. Monitoring & Troubleshooting
AWS Cloud9 IDE integration simplified debugging by providing direct terminal access to EC2 instances without managing SSH keys. CloudWatch logs would further enhance observability in production scenarios.
6. Cost Optimization Opportunities
Current architecture runs 24/7. Future optimization: implement Auto Scaling based on traffic patterns, use RDS instead of self-managed MariaDB, and configure instance scheduling for the development environment.
Project Details
Estimated Monthly Cost: <p class="whitespace-normal break-words"><strong>Development (us-east-1):</strong> ~$11/month</p> <ul class="[&:not(:last-child)_ul]:pb-1 [&:not(:last-child)_ol]:pb-1 list-disc space-y-2.5 pl-7"> <li class="whitespace-normal break-words">EC2 t2.micro, 8GB storage, data transfer, secrets management</li> </ul> <p class="whitespace-normal break-words"><strong>Production (us-west-2):</strong> ~$21/month</p> <ul class="[&:not(:last-child)_ul]:pb-1 [&:not(:last-child)_ol]:pb-1 list-disc space-y-2.5 pl-7"> <li class="whitespace-normal break-words">EC2 t2.small, 8GB storage, data transfer, secrets management</li> </ul> <p class="whitespace-normal break-words"><strong>Total:</strong> $32-35/month</p> <p class="whitespace-normal break-words"><strong>Optimizations Available:</strong></p> <ul class="[&:not(:last-child)_ul]:pb-1 [&:not(:last-child)_ol]:pb-1 list-disc space-y-2.5 pl-7"> <li class="whitespace-normal break-words">Reserved Instances: 40% savings</li> <li class="whitespace-normal break-words">RDS Free Tier: eliminates DB overhead</li> <li class="whitespace-normal break-words">Dev instance scheduling: 75% dev cost reduction</li> <li class="whitespace-normal break-words"><strong>Optimized Total: $15-20/month</strong></li> </ul>
Complexity: Intermediate
GitHub Repository: View on GitHub