Forem: Yogender

I Built a Self-Hosted Uptime Monitor in 1 Week Because UptimeRobot Only Lets You Monitor One Site for Free

Yogender — Sun, 05 Apr 2026 05:10:23 +0000

The Problem
I was building my project and needed to monitor its uptime. I went to UptimeRobot — and hit a wall. The free plan only lets you monitor one website. I had multiple services running and didn't want to pay just to know if my own server was down.
So I did what any developer does. I built it myself.

What I Built
Site-MOnitoring (Nexus) — a fully self-hosted, open-source uptime monitoring tool that:

📡 Monitors unlimited websites — add as many URLs as you want
📊 Shows real-time analytics — response times, status codes, uptime history
🎨 Interactive dashboard — built with React + Framer Motion
😴 Anti-sleep engine — keeps itself alive on free platforms like Render/Vercel
🏠 100% self-hosted — your data, your server, zero subscription fees

Live demo: pingbot.yogender1.me
GitHub: github.com/yogender-ai/Site-MOnitoring

Tech Stack
LayerTechnologyFrontendReact 18, Vite, TailwindCSS, Framer MotionBackendNode.js, Express.jsDatabasePostgreSQLJobsCron jobs (pinger.js)HostingRender (backend), Vercel (frontend)

The Hardest Part
Getting the anti-sleep engine right was tricky. Free-tier platforms like Render spin down your server after 15 minutes of inactivity. So your monitor — which is supposed to tell you when things go down — goes down itself.
I solved this by building a self-ping system using cron jobs that keeps the backend alive by pinging its own endpoint periodically. Simple idea, but took some debugging to get right without hammering the server.

Architecture
Nexus-Core/
├── backend/
│ ├── db.js # Database connection
│ ├── pinger.js # Background site poller + self-ping
│ └── server.js # Express API
└── frontend/
└── src/
├── api.js # Axios API calls
├── App.jsx # Main React app
└── index.css # TailwindCSS

What's Next
One valid criticism I've already received: single-server monitoring can give false positives. If your server goes down, it might report Google as down too. The fix is multi-region monitoring — checking from multiple servers in different locations.
That's on the roadmap. For now, Nexus is perfect for:

Monitoring your own projects and APIs
Self-hosted infrastructure
Anyone who wants full control without paying monthly

It's Open Source — Come Contribute!
I built this in 1 week and there's a lot more to add. I've opened up 4 beginner-friendly issues if you want to contribute:

🌙 Dark/light mode toggle
📧 Email alerts when sites go down
📄 CSV export for uptime logs
📝 CONTRIBUTING.md improvements

First time contributing to open source? This is a great project to start with. The codebase is small, clean, and well-documented.
⭐ If you find it useful, a star on GitHub helps a lot!
GitHub: github.com/yogender-ai/Site-MOnitoring

Built with React, Node.js, PostgreSQL, and a healthy frustration with paywalls.

Restarting DSA After a Long Break 🚀

Yogender — Sun, 05 Apr 2026 02:20:55 +0000

After a long pause… I’m finally back to Data Structures & Algorithms.

Life happened.
A lot happened.

There were distractions, changes, overthinking, maybe even burnout — and somewhere in between, consistency slipped away.

But today, I’m starting again. Not perfectly. Not with a grand plan. Just… starting.

📌 What’s different this time?

No pressure to be perfect
No unrealistic schedules
Just showing up daily
Focusing on understanding, not rushing

I’ve rebuilt my repo from scratch and will be pushing solutions consistently again. One problem at a time.

DSA isn’t just about cracking interviews — it’s about discipline, problem-solving, and proving to yourself that you can stay consistent even when it’s hard.

If you’re also restarting after a break, this is your sign:
👉 You’re not behind. You’re just getting started again.

Let’s do this. 💻🔥

🖐️ Hand-Controlled Particle Gravity Simulation (WebGL + Three.js)

Yogender — Sat, 31 Jan 2026 09:59:59 +0000

I built an interactive particle gravity simulation that reacts to real-time hand gestures — fully in the browser.

No game engine.
No plugins.
Just WebGL + Three.js + AI-based hand tracking.

You can literally turn your hand into a gravity well and control thousands of particles in real time.

🚀 Live Demo

👉 https://particle-gravity-fun.pages.dev/

💻 Source Code (Open Source)

👉 https://github.com/yogender-ai/Particle-Gravity-Fun

⭐ Feel free to star or fork if you find it interesting

✨ Features

~8,000 particles rendered in real time

Smooth 60 FPS particle swarming

Dynamic color physics (gesture-based color transitions)

Pinch gesture triggers high-energy “cosmic” effects

Runs entirely in the browser

🛠️ Tech Stack

Three.js (WebGL rendering)

JavaScript

AI Hand Tracking (MediaPipe)

WebAssembly (for performance-critical logic)

Cloudflare Pages (deployment)

🧠 Why I Built This

This started as a fun experiment to explore how far creative coding + AI + web graphics can go without heavy frameworks.

It’s exciting to see sci-fi-like interactions becoming possible directly on the web.

🙌 Feedback Welcome

I’m actively improving this project and experimenting with:

Better physics

More gesture types

Performance optimizations

Would love feedback from fellow developers 🙏
If you enjoyed it, please ⭐ the repo or share ideas.

🔖 Tags (important on dev.to)

Add these when publishing:

threejs #webgl #javascript #creativecoding #opensource #ai #frontend

KNN Algorithm from Scratch -Cat vs Dog Image Classification in Python (Complete Experiment)

Yogender — Tue, 06 Jan 2026 22:35:18 +0000

🧠 KNN Algorithm from Scratch — Real Image Classification Experiment

I recently built a K-Nearest Neighbors (KNN) algorithm from scratch in Python to perform cat vs dog image classification.

This project focuses on understanding how distance-based machine learning really works, rather than simply chasing accuracy.

🎯 What this project demonstrates

Machine Learning from scratch
KNN algorithm implementation
Image classification using raw pixel vectors
Supervised learning fundamentals
How noise, distance, and neighbors affect predictions
The curse of dimensionality in real experiments

🧪 The core experiment

After training the model, I discovered that removing only 5 noisy images from each class drastically improved prediction accuracy.

This happens because:

KNN has no abstraction or learning of features
All intelligence comes from the geometry of the data
A few bad samples can distort the entire decision boundary

🧬 Why this matters

Most tutorials skip the uncomfortable truth about KNN:

More data does not always mean better results
Distance metrics define the model’s intelligence
High-dimensional data behaves very unintuitively

Understanding this makes you a much stronger machine learning engineer.

📦 Complete Source Code

You can explore the full project here:

🔗 https://yogender-ai.github.io/knn-cat-dog-demo/

🏁 Final thought

Before neural networks learn to see,

we must understand how distance learns to lie.

If you’re learning machine learning, this experiment is worth exploring.