Overview

What began as a single server has grown into a fully‑featured private data center that I refer to as
MailoDataCenter. The goal is simple: gain complete control over a fleet of purpose‑built physical
machines that host game servers, websites, APIs, monitoring services, storage, and experimental
workloads—all powered by my own hardware.

---

Motivation

The Core Objective

• Full control – no reliance on third‑party cloud providers.
• High performance – raw hardware delivers the speed and responsiveness required for demanding
  workloads.
• Cost efficiency – a single infrastructure can run a wide range of services more cheaply than
  multiple VPS instances.

Server Roles

Server	Primary Function
Nebula	Heavy compute & virtualization
Nimbus	Storage and data services
Zenith	High‑performance storage & compute workloads

Each server is dedicated to a specific set of responsibilities but they all interconnect seamlessly to
form a unified environment.

---

Architecture & Stack

Layer	Technology	Purpose
Container Orchestration	Docker & Docker Compose	Simplified deployment & isolation
Platform‑as‑a‑Service	Coolify	Web‑based management panel
Ingress	Traefik	Dynamic reverse proxy, wildcard domains, automated HTTPS
Monitoring	Prometheus + Grafana + Node Exporter + cAdvisor	Metrics collection, visualization,
alerting
Status Pages	Uptime Kuma	Public & internal uptime dashboards
CI/CD	Self‑hosted GitHub runners	Automated builds & deployments
Custom Services	Node.js & other APIs	Tailored business logic & integrations

The container‑centric approach allows me to push a new repository to a GitHub branch and have the new
service spun up automatically.

---

Production Workloads

• Game Servers – Minecraft (Paper, Fabric, Velocity, modded) and other real‑time multiplayer
  platforms.
• Web & API Services – Personal sites, public APIs, and internal dashboards.
• Media & Storage – Large music libraries, media transcoding pipelines.
• Automation – Bots, scrapers, custom automation tools.
• Testing Environments – Rapid prototyping of new projects and features.
• Monitoring & Metrics – End‑to‑end visibility across all services.

Some game worlds exceed hundreds of gigabytes and require pre‑generated maps with 50 000+ chunks.

---

Storage & Performance Focus

Component	Configuration	Purpose
RAID Arrays	24 × 24 TB drives	High capacity, fault tolerance
NVMe Pools	3 × 1 TB SSDs	Low‑latency, high‑throughput workloads
SAS Arrays	8 × 6 TB HDDs	Bulk data archival
Networking	10 GbE interconnects	Minimal bottlenecks between nodes

All performance metrics are continuously measured, visualised in Grafana, and acted upon—if a component
slows, I identify, quantify, and remediate.

---

Why a Self‑Hosted Data Center?

Reason	Explanation
Cost	The scale of storage and compute required would be prohibitively expensive on public clouds.
Control	Full authority over configuration, security hardening, and optimisation.
Learning	Hands‑on exposure to Linux, networking, containerisation, and infrastructure
engineering.

The trade‑off is that I am the only “cloud provider” for my services, meaning any hardware failure or
software regression has immediate impact. However, this risk is an inherent part of operating a real data
center and is a valuable learning experience.

---

Long‑Term Vision

• Transform MailoDataCenter into a private cloud for my own projects.
• Potentially offer VPS or managed services to external clients.
• Continue evolving the platform into a scalable, production‑ready infrastructure playground.

---

Technical Demo: Server Status Simulator

Below is a simple Node.js script that simulates server load and status. It’s useful for testing
monitoring dashboards and alerting rules.

// Fake server status simulator

const servers = [
  { name: "Nebula", load: 0 },
  { name: "Nimbus", load: 0 },
  { name: "Zenith", load: 0 },
];

function randomLoad() {
  return Math.floor(Math.random() * 101); // 0 - 100%
}

function updateStatus() {
  console.clear();
  console.log("=== MailoDataCenter Status ===\n");

  servers.forEach(server => {
    server.load = randomLoad();

    let status = "OK";
    if (server.load > 85) status = "HIGH LOAD";
    if (server.load > 95) status = "CRITICAL";

    console.log(
      `${server.name.padEnd(8)} | Load: ${server.load}% | Status: ${status}`
    );
  });
}

// Update every 2 seconds
setInterval(updateStatus, 2000);

---

Closing Thoughts

Operating my own data center isn’t merely about reducing expenses. It’s about owning the entire stack,
achieving peak performance, and continuously learning. Seeing every component—from the physical servers
to the monitoring dashboards—run on my own hardware is profoundly rewarding and, admittedly, somewhat
addictive.