Load Balancers

When a website becomes popular, a single server can’t handle all the traffic. It’s like a restaurant with only one waiter—no matter how fast the waiter is, they’ll eventually get overwhelmed.

A Load Balancer is like a host at the front of the restaurant who directs customers to different waiters.

What is a Load Balancer?

A load balancer is a device or software that distributes incoming network traffic across multiple servers.

Examples: Nginx, HAProxy, AWS Elastic Load Balancer (ELB).

Architecture Diagram (The Load Balancer)

A load balancer sits in front of your server group:

       [ User (Client) ]
               ↓
       [ Load Balancer (Nginx) ]
               ↓
       ----------------- (Distributes traffic)
       ↓               ↓               ↓
   [ Server 1 ]    [ Server 2 ]    [ Server 3 ]
       ↓               ↓               ↓
       [----------------- Database -----------------]

Why Load Balancers Exist

1. High Availability

If one of your servers crashes, the load balancer simply stops sending traffic to it and redirects users to the other healthy servers. Your app stays online!

2. Scalability

You can easily add or remove servers as your traffic changes throughout the day.

3. Better Performance

By spreading the work across many servers, each server has more resources available to handle its requests quickly.

How Junior Developers Encounter This

Nginx: The most common tool for load balancing on a single server.
Auto-Scaling Groups: A cloud configuration that automatically adds more servers (and updates the load balancer) when your app gets busy.
Round Robin: The simplest way to balance load—sending one request to Server 1, the next to Server 2, and so on.

Common Beginner Mistakes

1. Sticky Sessions

If Server 1 stores a user’s session data and the next request goes to Server 2, the user will be “logged out.” You need to handle “sticky sessions” or store session data in a central database (like Redis).

2. Not Testing Failover

If you have three servers but your load balancer doesn’t know when one is down, users will get error messages 33% of the time! Always set up “Health Checks.”

3. Single Point of Failure

If you only have one load balancer, and it crashes, your whole app goes offline. In production, you often need two load balancers working together.

Real-World Example: Google Search

When you go to google.com, you aren’t talking to one single server. You’re talking to a massive load balancer that directs your search to one of thousands of servers across the world!

Next Lesson

Sometimes, you have tasks that take a long time to finish. You shouldn’t make the user wait:

➡ Message Queues