🛡️ DevOps Meets High Availability: Mastering Azure Availability Sets

In the fast-paced world of DevOps, we automate everything—from infrastructure provisioning to deployments and monitoring. But there’s one question that always haunts us:

💭 "What happens if one of my VMs goes down during a deployment or a patch?"

This is where Azure Availability Sets step in as unsung heroes of resilient infrastructure design.

Let’s explore what they are, why they matter in DevOps, and how to implement them in your automation pipelines.

🌐 What is an Azure Availability Set?

An Availability Set in Azure is a logical grouping of virtual machines (VMs) that helps protect your application from hardware failures and maintenance events.

Instead of placing all your VMs on the same physical host (a risky move), Azure distributes them across multiple Fault Domains and Update Domains.

🧱 Fault Domain vs. Update Domain — What's the Difference?

Here’s how Azure Availability Sets keep your VMs safe:

🧯 Fault Domain (FD): Represents a physical rack in the Azure datacenter. Each rack has its own power, cooling, and networking. By distributing VMs across different FDs, Azure ensures that a hardware failure doesn't take down all your VMs.

🛠️ Update Domain (UD): Represents a logical group of VMs that are updated (patched/rebooted) together during Azure’s planned maintenance. Azure updates one UD at a time to make sure some VMs stay available during system updates.

🔒 With at least 2 VMs in an Availability Set, Azure guarantees 99.95% SLA.

🧰 DevOps Use Case: Where It Fits In

As DevOps engineers, we often work with Terraform, Bicep, ARM templates, or Azure CLI in our CI/CD pipelines. Here’s how Availability Sets enhance your automation strategy:

✅ Resilient Deployments

Define your Availability Sets as Infrastructure as Code (IaC) and assign your VMs to them automatically in pipelines.

✅ Controlled Maintenance

When deploying production workloads, you can't afford a full outage during patching. Update Domains help maintain partial availability.

✅ Cost-Effective HA

You don’t need to scale out using complex services. Availability Sets give you high availability at no additional cost.

📦 Terraform Example

resource "azurerm_availability_set" "app_avset" {
  name                         = "web-avset"
  location                     = azurerm_resource_group.main.location
  resource_group_name          = azurerm_resource_group.main.name
  platform_fault_domain_count  = 2
  platform_update_domain_count = 5
  managed                      = true
  tags = {
    environment = "production"
  }
}
        

Add your VMs like this:

resource "azurerm_windows_virtual_machine" "web_vm" {
  name                = "web-vm1"
  availability_set_id = azurerm_availability_set.app_avset.id
  #add other arguments required for vm
}
        

🔁 Azure CLI in CI/CD (YAML)

- task: AzureCLI@2
  inputs:
    azureSubscription: 'AzureConnection'
    scriptType: 'bash'
    scriptLocation: 'inlineScript'
    inlineScript: |
      az vm create \
        --resource-group my-rg \
        --name myvm1 \
        --availability-set web-avset \
        --image UbuntuLTS \
        --admin-username azureuser \
        --generate-ssh-keys
        

🧠 When to Use Availability Sets

🔹 Hosting stateful apps (web, app, DB) on VMs 🔹 Lift-and-shift workloads where VMs are essential 🔹 Apps requiring 99.95% uptime without using Availability Zones or VM Scale Sets 🔹 VMs that don’t need autoscaling but must remain redundant

⚠️ Key Considerations

⚙️ VMs in an Availability Set must be in the same region and resource group ⚙️ Only applies to Azure VMs, not to App Services or containerized workloads ⚙️ Can't add existing VMs to an Availability Set—you must associate them at creation time.

🗺️ Availability Sets vs. Other Azure HA Options

🧵 Final Thoughts

In DevOps, we always say “design for failure”—because failure will happen.

Azure Availability Sets help ensure that even when things go wrong, your apps keep running. They’re simple to implement, cost nothing extra, and integrate beautifully with your CI/CD and IaC pipelines.

So the next time you're provisioning VMs in Azure, don’t just ask:

“Will it deploy?”

Ask:

✅ “Will it survive?”

💬 Your Turn Have you incorporated Availability Sets in your IaC or deployment pipelines? What other HA patterns do you use in Azure?

Let’s share knowledge and grow together! Drop your thoughts below ⬇️

#Azure #DevOps #Terraform #InfrastructureAsCode #AzureDevOps #HighAvailability #CloudEngineering #Automation #CI/CD #AzureVMs #ReliabilityEngineering #CloudBestPractices