Platform Operational
Cloud Support Platform on AWS
A production-style cloud environment built to demonstrate
infrastructure automation, Linux administration, observability,
incident response, and technical troubleshooting.
The Flask application runs on Amazon EC2 behind NGINX and
Gunicorn, is secured with HTTPS, monitored through Amazon
CloudWatch, and provisioned using Terraform.
What I Built
- Provisioned AWS infrastructure using Terraform
- Deployed a Flask application on Amazon EC2
- Configured NGINX as a reverse proxy for Gunicorn
- Secured the application using HTTPS and TLS certificates
- Removed public SSH access and enabled AWS Systems Manager Session Manager
- Configured CloudWatch log collection, alarms, and SNS notifications
- Automated CloudWatch Agent installation through EC2 user data
- Migrated the production domain to the Terraform-managed server
- Created least-privilege IAM access for log analysis
Breaking, Fixing and Troubleshooting
This project was not only about deploying infrastructure.
I deliberately created and investigated failures across the
platform to practise real production-support workflows.
- Diagnosed NGINX and Gunicorn service failures
- Investigated and resolved 502 Bad Gateway errors
- Troubleshot DNS, HTTPS, security-group, and IAM issues
- Reviewed Linux, systemd, NGINX, and CloudWatch logs
- Validated ports, processes, application health, and network reachability
- Tested CloudWatch alarms and confirmed SNS recovery notifications
- Identified automated internet scanning and suspicious HTTP requests
- Documented root causes, remediation steps, and customer-facing updates
AI Log Intelligence
The platform includes a Python-based AI log intelligence workflow
that retrieves operational events from Amazon CloudWatch and
analyzes them using the OpenAI API.
The workflow helps produce:
- Request, success, and failure counts
- HTTP error and traffic-pattern summaries
- Likely root-cause hypotheses
- Incident severity classifications
- Recommended troubleshooting and remediation steps
- AWS and Linux diagnostic commands
- Structured incident reports
- Customer-ready incident updates
AI is used to accelerate investigation and summarize evidence,
while final conclusions are validated against the original logs.
Tools and Technologies
AWS EC2
Terraform
Linux
NGINX
Gunicorn
Flask
Python
Bash
Amazon CloudWatch
Amazon SNS
AWS IAM
Systems Manager
Elastic IP
DNS
HTTPS/TLS
Let's Encrypt
Certbot
systemd
boto3
OpenAI API
GitHub
Current Project Stage
The environment is currently operational, publicly accessible,
aligned with its Terraform configuration, and monitored through
CloudWatch.
Current work is focused on improving deployment automation,
documentation, health checks, monitoring coverage, and reducing
the remaining dependency on the reusable application AMI.
View health endpoint
|
View skills endpoint