All articles
Azure AI27 Aug 2026 · 2 min read

What Azure Load Balancer Does and When You Actually Need One

A plain-language look at spreading traffic across VMs without the networking jargon

By Shehryar Hassan, Microsoft 365 & Azure Consultant

Someone asked me last week why their app kept timing out even though the VM behind it looked fine in the portal. Turned out they had two VMs running the same web app, but nothing was splitting traffic between them, so one VM was getting hammered while the other sat idle. That's the exact problem Azure Load Balancer solves.

What it actually does

Azure Load Balancer sits in front of a group of VMs and spreads incoming traffic across them based on simple rules, usually round robin plus a health check. If one VM stops responding, the load balancer notices within seconds and quietly stops sending it traffic until it recovers. Nobody has to click anything.

There are two flavors. A public load balancer takes traffic from the internet and distributes it inward, useful for anything customer facing. An internal load balancer only handles traffic inside your virtual network, which is what you want for backend services like a database tier or an internal API that other apps in Azure talk to but the public never sees.

When you actually need one

If you're running a single VM, you don't need this at all, there's nothing to balance. The moment you add a second VM for the same workload, for redundancy or to handle more load, that's when a load balancer earns its keep. It's also worth setting one up before you need it, not after an outage teaches you the hard way that one VM going down took your whole app with it.

One thing that trips people up: Azure Load Balancer works at the network level, so it doesn't look at the content of the request, just where it's going. If you need routing based on URL paths or SSL termination, you actually want Application Gateway instead, they solve related but different problems.

Setting one up isn't complicated. You define a frontend IP, a backend pool of VMs, a health probe, and a load balancing rule that ties it together. The health probe is the part worth getting right, if it's checking the wrong port or path, the load balancer might send traffic to a VM that's actually broken even though the VM itself looks healthy.

#Azure#Load Balancer

Get new posts by email

One note when there is something worth reading. No spam, unsubscribe anytime.

By subscribing you agree to the privacy policy.