Azure Regions and Availability Zones: How They Protect Your App

AZ-900 · Updated June 2026

When you deploy something to Azure, you choose where it runs. That choice determines how resilient your application is to hardware failure, power outages, and even large-scale disasters. Azure regions and availability zones are the two building blocks that make high availability possible , and understanding how they work is a core requirement for the AZ-900 exam.

What an Azure region is

An Azure region is a geographical area containing one or more datacentres that are connected through a low-latency network. When you create a virtual machine, a storage account, or a database, you pick a region , West Europe, East US, Southeast Asia, and so on. All the resources you deploy in that region are physically close to each other and share the same underlying network infrastructure.

Azure has more than 60 regions worldwide, spread across the Americas, Europe, Asia Pacific, Africa, and the Middle East. The region you choose affects latency (closer to your users is faster), data residency (some regulations require data to stay in a specific country), and which services are available (not every service is offered in every region).

Some Azure regions are paired with another region in the same geography. These are called region pairs, and Microsoft uses them to coordinate planned maintenance and replicate certain services. For example, West Europe is paired with North Europe. If Microsoft needs to update infrastructure across both regions, they do it one at a time so at least one is always available.

What availability zones are

Within a region, Azure divides infrastructure into availability zones. Each zone is a physically separate datacenter , its own building, with its own independent power supply, cooling system, and network connections. Zones within the same region are connected by high-speed, low-latency private links, but a failure in one zone (a power fault, a cooling failure, a network issue) does not affect the others.

Most Azure regions with availability zone support have three zones, labelled Zone 1, Zone 2, and Zone 3. When you deploy across all three zones, your application can survive the complete loss of one zone and keep running without interruption.

Zonal vs zone-redundant services

Azure services that support availability zones fall into two categories. Understanding the difference matters for both exam questions and real deployments.

• arrow_rightZonal services let you pin a resource to a specific zone. A virtual machine deployed to Zone 1 runs only in that zone. If Zone 1 fails, that VM goes down. To get resilience, you deploy multiple VMs across different zones yourself.
• arrow_rightZone-redundant services automatically spread data and compute across all three zones. Azure manages the distribution for you. If one zone fails, the service continues running from the other two with no action required from you. Azure SQL Database with zone redundancy and zone-redundant storage (ZRS) work this way.

The geography layer above regions

Azure organises regions into geographies , typically aligned with a country or group of countries. A geography contains at least one region, often two (a region pair). The geography boundary matters for data residency: data stored in a geography stays within that geography unless you explicitly configure cross-geography replication.

The full hierarchy from largest to smallest is: geography → region pair → region → availability zone → datacenter. AZ-900 expects you to know this hierarchy and where each level sits.

How availability zones connect to SLAs

Deploying across availability zones directly affects the SLA Microsoft can offer. A single virtual machine using Premium SSD storage has a 99.9% uptime SLA. The same VM deployed across two availability zones gets a 99.99% SLA. That small percentage difference translates to significantly less allowable downtime per year.

This is the mechanism behind high availability in Azure: spreading workloads across physically isolated zones means no single point of hardware failure can take down your application. The more isolation you build in, the stronger the SLA.

How to answer region and availability zone questions on the exam

AZ-900 uses regions and availability zones in two main question patterns: resilience scenarios and service placement scenarios.

1. 1

   Resilience scenarios

   A company wants an application that remains available if a single datacenter fails. The answer is availability zones , deploy across multiple zones within the same region. A company wants protection against a full regional outage. The answer is region pairs or deploying across two separate regions.

2. 2

   Service placement scenarios

   A question asks where to deploy a resource to minimise latency for users in Germany. The answer is a region in Germany or Western Europe. A question asks which feature separates datacentres within a region for fault isolation. The answer is availability zones.

3. 3

   Definition questions

   Know that a region is a set of datacentres within a latency perimeter. Know that availability zones are physically separate datacentres within a region. Know that region pairs are two regions in the same geography used for disaster recovery and planned maintenance coordination.

Frequently asked questions

Regions and availability zones are not just infrastructure concepts , they are the decisions that determine whether your application survives failure. On the AZ-900 exam, every question about high availability, fault tolerance, or disaster recovery will connect back to these two building blocks.

The mental model to carry into the exam: regions separate workloads geographically, availability zones separate them physically within a region. More separation means more resilience, and more resilience means a stronger SLA.