Azure Has Five Load Balancers — Here’s How to Pick

How to choose the right Azure entry point without redesigning your platform six months later

Motivation

A pattern I keep seeing in Azure reviews is this: the architecture diagram looks clean, the service choices look reasonable, and then the first failover test exposes a hidden mismatch.

Sometimes it’s DNS-based routing where the team expected proxy behavior. Sometimes it’s a pure L4 choice for a workload that really needed L7 controls. Sometimes it’s a regional service solving a global entry problem.

The root cause is usually the same: the routing model was never explicitly decided.

Azure doesn’t have “too many” load balancers. It has multiple entry points because the routing models are genuinely different.

Think of this as a decision guide for Azure engineers and architects who want to choose once and avoid expensive rewrites.

Quick answer: Decide global vs regional first, then L4 vs L7. Everything else follows from those two choices.

What This Article Does NOT Cover

This guide focuses on the five general-purpose load-balancing services. It does not cover Azure API Management’s built-in load balancing, Azure Kubernetes ingress…