What is a load balancer and when do you actually need one?
Tutorials always draw a load balancer box in architecture diagrams, but my side project runs fine on one server. At what point does this become real?
132
2 answers2.9K views2 Answers
Accepted answer
A load balancer is a traffic cop: it sits in front of 2+ identical servers and spreads incoming requests across them, so no single machine drowns and a dead server stops receiving traffic.
You need one when either: (a) one server can't handle your traffic anymore — you've scaled vertically (bigger machine) as far as is sensible, or (b) you can't afford downtime — two servers behind a balancer means one can die or be redeployed while the other serves.
For a side project? You almost certainly have one already without knowing — Vercel, Railway, and every PaaS run balancers in front of your app. You only manage your own when you leave managed platforms, typically well past 10k+ concurrent users territory.
34
Interview tip: the follow-up is always 'how does the balancer decide?' Know round-robin, least-connections, and IP hashing, and mention health checks — that it stops routing to instances that fail them. Those four terms cover 90% of the question.
26
Related questions
- Best free resources to learn system design in 2026?1 answer
- Monolith vs microservices for a small team — what's the honest answer?1 answer
- Horizontal vs vertical scaling — what's the difference and which comes first?2 answers
- How does caching work and what should I cache first?1 answer
- What is the difference between SQL and NoSQL databases, and when do I choose each?1 answer
- How do I start learning system design as a fresher?1 answer