Campus Server Inventory & Web Observability Stack

Hey everyone If you are a student too you definitely know the struggle when campus websites crash during course registration or exam results day I was curious how real IT admins handle this and keep the servers alive So I built this project to learn how to setup and monitor a production-ready web infrastructure

The Problem & How I Solved It

The Problem: Usually beginners put everything (web frontend and database) in one single machine If one thing breaks the whole campus web goes down completely Also putting your database in public is a massive security risk

My Solution: I used Docker Compose to split the system into isolated microservices:

• The Storefront (Apache2): Handles visitor traffic on port 8080
• The Vault (MySQL Database): Hidden deeply inside a private virtual network to keep campus asset data safe
• The Watchman (Prometheus + Blackbox Exporter): Constantly pings the web server to check if it is responding
• The Command Center (Grafana): A dashboard that turns raw data into clear ONLINE or OFFLINE status

System Architecture & Data Flow

Here is how all the pieces fit together inside the host machine

┌────────────────────────────────────────────────────────────────────────┐
│                          LOCAL HOST MACHINE                            │
│                       (Docker Desktop / WSL2)                          │
│                                                                        │
│   Student / Admin Access                                               │
│   http://localhost:8080                                                │
│        │                                                               │
│   ┌────▼───────────────────────────────────────────────────────────┐   │
│   │             ISOLATED DOCKER BRIDGE (`web_network`)             │   │
│   │                                                                │   │
│   │  ┌────────────────────────┐     ┌──────────────────────────┐   │   │
│   │  │   Apache2 Web Server   │     │   MySQL Database Engine  │   │   │
│   │  │   (Container: apache_web)│   │   (Container: mysql_db)  │   │   │
│   │  │                        │     │                          │   │   │
│   │  │   Port: 8080 -> 80     │     │   Port: 3306 (Internal)  │   │   │
│   │  │   Serves the Frontend  │     │   Holds Campus Assets    │   │   │
│   │  └───────────▲────────────┘     └─────────────▲────────────┘   │   │
│   │              │                                │                │   │
│   │         HTTP Pings                       SQL Queries           │   │
│   │              │                                │                │   │
│   │  ┌───────────┴────────────┐                   │                │   │
│   │  │    Blackbox Exporter   │                   │                │   │
│   │  └───────────▲────────────┘                   │                │   │
│   │              │ Metrics Scrape                 │                │   │
│   │  ┌───────────┴────────────┐                   │                │   │
│   │  │       Prometheus       │───────────────────┘                │   │
│   │  └───────────▲────────────┘                                    │   │
│   │              │ PromQL                                          │   │
│   │  ┌───────────┴────────────┐                                    │   │
│   │  │    Grafana Dashboard   │ ──▶ Opens at http://localhost:3000 │   │
│   │  └────────────────────────┘                                    │   │
│   └────────────────────────────────────────────────────────────────┘   │
└────────────────────────────────────────────────────────────────────────┘

My Troubleshooting Journey

Building this was fun but challenging I ran into a couple of issues that forced me to learn how real SysAdmins work

• The Database Security Trap: At first I used basic passwords like securepassword123 and a user named campus_admin I realized that is bad for real-world security so I upgraded the infrastructure to use strict enterprise-style credentials like infra_db_admin
• The Web Monitoring Dilemma: Monitoring a live web application is different from monitoring a router I had to learn how to configure the Blackbox Exporter to send internal HTTP requests and map those outputs into Grafana so it shows a simple text status

Project Directory Setup

I keep my code clean and structured like this

web-infra-stack/
│
├── docker-compose.yml          # The blueprint orchestrating all 5 containers
├── init.sql                    # Auto-seeding script for our campus database
├── html/                       
│   └── index.html              # Simple frontend landing page
└── prometheus/
    └── prometheus.yml          # Scrape targets and Blackbox routing rules

Testing the Stack (Proof of Concept)

I tested this to make sure the automation works perfectly without manual intervention

1 Database Automation Test

When the MySQL container boots up it automatically reads init.sql to build the database. Here is the exact schema running under the hood:

CREATE DATABASE IF NOT EXISTS campus_info_db;
USE campus_info_db;

CREATE TABLE IF NOT EXISTS server_inventory (
    id INT AUTO_INCREMENT PRIMARY KEY,
    server_name VARCHAR(100) NOT NULL,
    ip_address VARCHAR(15) NOT NULL,
    building_location VARCHAR(100) NOT NULL,
    managing_department VARCHAR(100) NOT NULL,
    primary_function TEXT NOT NULL,
    status ENUM('ONLINE', 'MAINTENANCE', 'OFFLINE') DEFAULT 'ONLINE',
    last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
);

I ran a terminal check inside the live container to verify the seeding process:

docker exec -it mysql_db mysql -u infra_db_admin -p"SysP@ssW0rd!2026" campus_info_db -e "SELECT * FROM server_inventory;"

The Result: The terminal perfectly printed a structured asset management table containing my campus server mock data

2 Live Outage Simulation Test

I wanted to test if my Grafana dashboard is smart enough to flag an emergency I intentionally killed the web server container using the terminal:

docker compose stop web

The Reaction: Within seconds the Blackbox exporter noticed the failure and Prometheus caught it My Grafana dashboard dynamically flipped its massive green ONLINE banner into a bright red OFFLINE warning just like this:

Running docker compose start web brought everything back to normal instantly It flows like water and shows exactly how the monitoring reacts to a real problem

Conclusion

This project taught me that managing an IT infrastructure is not just about writing code It is about automation and building reliable foundations Being able to spin up a web server and a fully functional monitoring pipeline with just one command docker compose up -d feels like absolute magic Thanks for checking out my repository