AWS migration and CI/CD

The challenge

A healthcare provider ran a PHP monolith on a single unmanaged server. Releases were manual, downtime was routine and there was no tested recovery path.

What we did

• Containerised the application without a risky rewrite.
• Described all infrastructure in Terraform with separate staging and production.
• Built a CI/CD pipeline with automated tests, migrations and rollback.
• Added monitoring, alerting and automated encrypted backups.

The outcome

• Release cadence moved from monthly to daily.
• Recovery tested and documented, measured in minutes.
• Hosting cost reduced despite higher availability.

Stack

• AWS ECS
• Docker
• Terraform
• GitHub Actions
• RDS