What I built, and what came of it.
Clients and their sectors are not published. The figures are — they are the ones there were.
- 01
From AWS to two Proxmox VPS
$500 → $120 a month, with more database redundancy than before.
A Laravel application with three environments and around a hundred recurring users was paying for the structure of an established platform. Two independent nodes, live MySQL replication and a Cloudflare load balancer in front. A return to AWS is planned for when the audience justifies it.
Head of DevOps · design and delivery
- Proxmox
- MySQL
- Cloudflare
- Laravel
- 02
Infrastructure for 3,000-user peaks
Holds 3,000 people arriving in the same minute.
A Laravel web app on AWS, with all infrastructure defined as code. High availability across multiple availability zones, rate limiting and SQL injection protection on the ingress path.
Head of DevOps · design and delivery
- AWS
- Terraform
- Multi-AZ
- Laravel
- 03
Rightsizing and cost attribution on AWS
Spend came down, and for the first time it was possible to say which team spends what.
A development team was over-provisioning and nobody could attribute the bill. Rightsizing of the oversized resources, tagging by team so every resource has an owner, and budget alerts.
Head of DevOps · design and delivery
- AWS
- FinOps
- Cost Allocation Tags
- Budgets
- 04
Migrating a geospatial platform to AWS
Bare metal to AWS in a one-hour window, and the code outlasted the handover.
A Java map application with a redundancy requirement on location data. Bastion, EC2, multi-AZ RDS, secrets management, Route 53 and a metrics dashboard with alerts, all in Terraform. Another DevOps engineer picked up that repository and kept working on it.
DevOps · implementation team
- AWS
- Terraform
- RDS Multi-AZ
- Route 53