The temptation to self-host a load balancer

When you need to distribute traffic across multiple servers, the instinct is to set up your own load balancer. HAProxy, Nginx, or Traefik on a dedicated VPS.

It works. But should you do it?

The problem with self-hosted load balancers

Adding a load balancer means adding another layer of infrastructure:

Another server to maintain

Security patches and updates
Monitoring and alerting
Backup and disaster recovery
SSL certificate management

Your load balancer becomes a critical single point of failure. If it goes down, everything behind it is unreachable.

High availability requires more complexity

A single load balancer defeats the purpose. You need:

Two load balancers for redundancy
Keepalived or similar for failover
Floating IPs or DNS failover

Now you’re maintaining multiple servers just to route traffic.

Geographic limitations

A self-hosted load balancer sits in one location. Users on the other side of the world route through that single point, adding latency.

The Cloudflare alternative

Cloudflare offers load balancing as a service. For a fraction of what you’d spend on load balancer infrastructure, you get:

Global anycast network

Cloudflare has data centers in over 300 cities worldwide. Your users connect to the nearest one automatically. No configuration required.

Health checks and automatic failover

Cloudflare monitors your origin servers and routes around failures automatically. If a server goes down, traffic shifts to healthy servers within seconds.

Geographic steering

Route users to the server closest to them:

European users → Amsterdam server
US users → Las Vegas or Dallas server
Asian users → Singapore server

This happens automatically based on user location.

Simple configuration

No HAProxy config files. No Nginx upstream blocks. Just a web dashboard where you:

Add your origin servers
Configure health checks
Set routing rules

That’s it.

What it costs

Cloudflare’s load balancing pricing is straightforward:

Load Balancing: Starts at $5/month for basic features
Additional origins: Small additional cost per origin server
Health checks: Included

Compare that to:

Two VPS instances for redundant load balancers: $10-40/month
Your time maintaining them: priceless (or expensive, depending how you look at it)

How we use it at ServerPoint

We practice what we preach. Our own infrastructure uses Cloudflare’s global load balancer:

ServerPoint’s Client Portal: Load balanced across multiple backend servers
APIs: Geographic routing to the nearest data center

When you log into portal.serverpoint.com, Cloudflare routes you to the best available backend automatically.

Setting it up

Step 1: Add your domain to Cloudflare

If you haven’t already, add your domain to Cloudflare and point your nameservers to them.

Step 2: Create a load balancer

In the Cloudflare dashboard:

Go to Traffic → Load Balancing
Click Create Load Balancer
Enter the hostname (e.g., app.yourdomain.com)

Step 3: Create origin pools

Origin pools are groups of servers. Create one or more:

Example: Single pool with multiple servers

Pool name: web-servers
Origins:
- server1.yourdomain.com (your first VPS)
- server2.yourdomain.com (your second VPS)

Example: Geographic pools

Pool us-servers: Dallas and Las Vegas VPS
Pool eu-servers: Amsterdam VPS
Pool asia-servers: Singapore VPS

Step 4: Configure health checks

Set up health checks to monitor your servers:

Type: HTTP or HTTPS
Path: /health or any endpoint that returns 200 OK
Interval: How often to check (60 seconds is common)

Create a simple health endpoint in your application:

app.get('/health', (req, res) => {
  res.status(200).send('OK')
})

Step 5: Set traffic steering

Choose how traffic is distributed:

Off: Random distribution
Geo Steering: Route to the nearest pool based on user location
Dynamic Steering: Route based on latency measurements
Proximity Steering: Route to the geographically closest pool

For most setups, Geo Steering or Proximity Steering works best.

Step 6: Configure fallback

Set a fallback pool in case your primary pools are all unhealthy. This ensures traffic always has somewhere to go.

What about SSL?

Cloudflare handles SSL termination at their edge. You can:

Use Cloudflare’s free SSL certificate for the frontend
Use origin certificates for the connection between Cloudflare and your servers

No more managing Let’s Encrypt renewals on a load balancer.

When you might still want self-hosted

There are edge cases where self-hosted load balancing makes sense:

Internal traffic that shouldn’t route through Cloudflare
Non-HTTP protocols that Cloudflare doesn’t support
Ultra-low latency requirements where every millisecond matters
Regulatory requirements that prohibit third-party routing

For most web applications and APIs, Cloudflare handles it better than you could yourself.

The bottom line

Every layer of infrastructure you add is a layer you have to maintain. Load balancers are critical infrastructure. If they fail, everything fails.

Cloudflare gives you:

Global presence you couldn’t build yourself
Automatic failover without complex configuration
Geographic routing out of the box
Less to maintain so you can focus on your application

For a few dollars a month, you get enterprise-grade load balancing without the enterprise-grade complexity.

Don’t build what you can buy cheaply, especially when the bought solution is better than what you’d build.

Explore our VPS plans across multiple data centers for geographic redundancy.

Skip the self-hosted load balancer, use Cloudflare instead