← all projects
Infrastructure 2023 – present

Self-Hosted Production Server

A multi-application production server on repurposed hardware — 8+ containerized services, SSL, isolated networks, automated backups, 9+ months zero unplanned downtime.

9+ mo, 0 down
uptime (unplanned)
8+
services in prod
backups / day
~$0 + power
monthly cost

// problem

Paying clients needed reliable hosting for a growing set of apps, but managed PaaS pricing scaled badly for small Sri Lankan businesses and gave me no control over the network, SSL, or backups. I needed production-grade reliability on a hobbyist budget.

// approach

I repurposed a desktop PC on a fiber line, installed Ubuntu Server, and manage it entirely over SSH. Every service is a Docker container on an isolated network, fronted by Nginx Proxy Manager for per-subdomain routing and automatic Let's Encrypt SSL.

Reliability is engineered, not hoped for: 3× daily backup cron jobs ship data off-box, containers carry restart policies, and the only downtime is a weekly scheduled restart. Portainer and n8n give me a dashboard and automation layer over the whole fleet.

// outcome

The box has run 8+ containerized services with zero unplanned downtime across 9+ months. Owning the full stack — hardware, network, SSL, backups — turned deployment concerns into design inputs and made me a sharper backend engineer.