There is a weird obsession in the dev community where people try to design complex architectures for millions of users before they even have five actual people using their app.
If you’re building a side project, a basic SaaS tool, or a portfolio site, you don't need microservices or an insane cloud setup. You need a simple, clean setup that runs fast and works. Overengineering a project before it has any users is usually just procrastination disguised as "scalability."
Keep it simple until it actually breaks. What’s the simplest, most basic tech stack you’ve used that successfully worked or got real users?
Website: ahmershah.dev
Medium: @syedahmershah
YouTube: @ahmershahdev
If Twitter could start as a fragile Ruby on Rails app that crashed constantly (the famous Fail Whale), your weekend SaaS project will survive on a basic digital ocean droplet.
This mindset shift changed everything for me. Now I build everything as a boring monolith. It lets me launch in days instead of months.
SQLite is severely underrated for hobby projects. It can easily handle hundreds of thousands of requests a day on a cheap server. Keep it simple!
Building for scale before you have traction is just a form of procrastination. It’s easier to solve technical scaling problems in your head than to solve product-market fit in the real world
Reminds me of the joke: Dev spends 3 months building a globally distributed, multi-region database fallback for an app that has 12 registered users.
Over-engineering kills the fun of hobby projects. You get so bogged down in setting up infrastructure, caching layers, and message queues that you never actually build the core feature.
The 'Next.js + Vercel + Postgres' stack handles an incredible amount of traffic out of the box. There’s absolutely no reason a side project needs microservices from day one.
Louder for the people in the back! Just use a simple VPS, a single monolith repo, and a relational database. If you actually hit a million users, you'll have the money and the motivation to rewrite it.
Premature optimization is the root of all uncompleted hobby projects! I spent two weeks setting up Kubernetes for a site that gets 5 visits a month, mostly from my mom.
Vijay Bhalae
On the flip side, sometimes people over-engineer hobby projects precisely because they want to learn those scaling tools for their resume. But if the goal is to launch a product, keep it simple.