Availability

Lab Objective

In this hands-on lab, you will learn how to:

Deploy VMs to availability zones across three separate datacenters for maximum resilience
Create availability sets with fault and update domains for traditional high availability
Configure load balancers to distribute traffic across multiple VM instances
Test failover scenarios by stopping VMs and observing traffic redirection
Set up basic monitoring to track VM availability and performance

Scenario: Deploy a web application across multiple availability zones with automatic failover to ensure high availability.

Pre-Provisioned Environment

Resource Type	Resource Name	Purpose
Resource Group	`Availability-Lab-RG`	Contains all lab resources
Virtual Network	`availability-vnet`	Network for VMs (10.0.0.0/16)
Load Balancer	`avail-lb[unique]`	Distributes traffic across VMs
Public IP	`avail-pip[unique]`	External access to load balancer

Lab Exercises

Part 1: Deploy VMs to Availability Zones

Step 1: Create First VM in Zone 1

Navigate to Virtual machines → Create
Select Availability-Lab-RG resource group
Set VM name: WebVM-Zone1
Choose Windows Server 2019 Datacenter
Select size: Standard_B2s
Set administrator username: azureuser
Set administrator password: Password123!

Step 2: Configure Availability Zone

Click “Availability options”
Select “Availability zone”
Choose “Zone 1”
Click “Next: Disks”

Step 3: Configure Networking

Select virtual network: availability-vnet
Create new subnet: zone1-subnet (10.0.1.0/24)
Select “Create new” public IP
Click “Review + create” → Create

Step 4: Create VMs in Other Zones

Repeat the process to create:
- WebVM-Zone2 in Availability Zone 2 (subnet: 10.0.2.0/24)
- WebVM-Zone3 in Availability Zone 3 (subnet: 10.0.3.0/24)

Verify: Check that all three VMs are created in different availability zones.

Part 2: Configure Load Balancer

Step 1: Add VMs to Backend Pool

Navigate to Load balancers → avail-lb[unique]
Go to Settings → Backend pools
Click the existing backend pool
Add all three VMs (WebVM-Zone1, WebVM-Zone2, WebVM-Zone3)
Click “Save”

Step 2: Configure Health Probe

Go to Settings → Health probes
Click “Add”
Set name: web-health-probe
Set protocol: HTTP
Set port: 80
Set path: /
Click “OK”

Step 3: Create Load Balancing Rule

Go to Settings → Load balancing rules
Click “Add”
Set name: web-lb-rule
Set frontend port: 80
Set backend port: 80
Select health probe: web-health-probe
Click “OK”

Verify: Load balancer is configured with all VMs in backend pool.

Part 3: Install Web Server on VMs

Step 1: Connect to First VM

Go to Virtual machines → WebVM-Zone1
Click “Connect” → “RDP”
Download RDP file and connect using azureuser / Password123!

Step 2: Install IIS

Open Server Manager
Click “Add roles and features”
Select “Web Server (IIS)” role
Complete installation
Open browser and verify IIS default page loads

Step 3: Customize Web Page

Navigate to C:\inetpub\wwwroot
Edit iisstart.htm
Add text: “This is WebVM-Zone1”
Save the file

Step 4: Repeat for Other VMs

Connect to WebVM-Zone2 and WebVM-Zone3
Install IIS on both VMs
Customize pages to show “WebVM-Zone2” and “WebVM-Zone3”

Verify: Each VM shows its unique identifier when accessed directly.

Part 4: Test Load Balancer

Step 1: Access Load Balancer IP

Go to Load balancers → avail-lb[unique]
Copy the public IP address
Open browser and navigate to the IP address
Refresh page multiple times

Expected: You should see different VM pages as load balancer distributes traffic.

Step 2: Test Failover

Stop one VM (WebVM-Zone1)
Refresh browser page multiple times
Observe that traffic only goes to remaining two VMs

Step 3: Restart VM and Verify

Start WebVM-Zone1 again
Wait 2-3 minutes for health probe to detect
Refresh browser to see all three VMs responding again

Verify: Load balancer automatically removes failed VMs and adds them back when healthy.

Part 5: Create Availability Set

Step 1: Create Availability Set

Navigate to Availability sets → Create
Set name: web-availability-set
Select resource group: Availability-Lab-RG
Set fault domains: 2
Set update domains: 5
Click “Create”

Step 2: Create VM in Availability Set

Create new VM: AvailSetVM1
Choose availability options: “Availability set”
Select web-availability-set
Use same credentials and settings as before
Create the VM

Step 3: Add Second VM to Set

Create another VM: AvailSetVM2
Select same availability set: web-availability-set
Complete creation

Verify: Both VMs are in the availability set with different fault domains.

Part 6: Monitor Availability

Step 1: Enable VM Insights

Go to Virtual machines → WebVM-Zone1
Click “Insights” under Monitoring
Click “Enable”
Select Log Analytics workspace or create new
Wait for deployment to complete

Step 2: View Availability Metrics

Go to Monitor → Metrics
Select resource: WebVM-Zone1
Add metric: “VM Availability”
Set time range: Last 4 hours

Step 3: Create Simple Alert

Go to Monitor → Alerts
Click “Create alert rule”
Select resource: WebVM-Zone1
Add condition: VM Availability less than 1
Create action group with email notification
Name the alert: “VM Down Alert”
Create the alert

Verify: Alert is created and will notify when VM becomes unavailable.

Troubleshooting Guide

Common Availability Issues

Issue	Symptoms	Possible Cause	Solution
VMs not distributing across zones	All instances in single zone despite configuration	Capacity constraints or SKU limitations	Verify VM size availability in all zones and consider different SKU
Load balancer not detecting failures	Traffic continues to unhealthy instances	Health probe misconfiguration or endpoint issues	Check health probe settings and ensure application health endpoint responds
Availability set deployment failures	Cannot add VM to existing availability set	Mixing managed and unmanaged disks	Ensure consistent disk types (managed disks recommended)
Zone-redundant services unavailable	Options grayed out during resource creation	Region or service limitations	Verify regional support for zone-redundant services and select supported region
High latency between zones	Poor application performance across zones	Cross-zone network latency or configuration	Optimize application architecture for zone-aware design and minimize cross-zone traffic

Availability Configuration Checklist

Component	Requirement	Status	Notes
VM placement	Distributed across zones or fault domains	✅	Ensures protection against infrastructure failures
Load balancer configuration	Health probes and backend pools configured	✅	Enables automatic traffic distribution and failover
Storage resilience	Zone-redundant or geo-redundant storage	✅	Protects against data loss during infrastructure failures
Network redundancy	Multiple paths and zone-redundant components	✅	Eliminates network single points of failure
Monitoring and alerting	Availability metrics and alerts configured	✅	Provides visibility and proactive issue detection

Best Practices

Scenario	Recommendation	Benefit
Mission-critical applications	Deploy across availability zones with zone-redundant services	Provides highest level of availability (99.99% SLA)
Cost-sensitive workloads	Use availability sets for basic high availability	Balances cost and availability for standard applications
Stateful applications	Implement data replication across zones or regions	Ensures data consistency and availability during failures
Global applications	Deploy across multiple regions with traffic manager	Provides geographic redundancy and disaster recovery
Maintenance planning	Align update domains with maintenance windows	Minimizes business impact during planned maintenance

Optional Advanced Exercises

For users wanting more technical depth, try these exercises:

Advanced Exercise 1: Automated Availability Deployment

# Deploy high availability infrastructure using PowerShell automation
Connect-AzAccount

# Set up variables for multi-zone deployment
$resourceGroup = "Availability-Lab-RG"
$location = "East US"
$vmSize = "Standard_B2s"
$availabilityZones = @("1", "2", "3")

# Create availability zones VMs with automation
foreach ($zone in $availabilityZones) {
    $vmName = "AutoVM-Zone$zone"
    $subnetName = "Zone${zone}Subnet"

    # Create VM configuration
    $vmConfig = New-AzVMConfig -VMName $vmName -VMSize $vmSize -Zone $zone
    $vmConfig = Set-AzVMOperatingSystem -VM $vmConfig -Windows -ComputerName $vmName -Credential (Get-Credential)
    $vmConfig = Set-AzVMSourceImage -VM $vmConfig -PublisherName "MicrosoftWindowsServer" -Offer "WindowsServer" -Skus "2019-Datacenter" -Version "latest"

    # Create VM with zone placement
    New-AzVM -ResourceGroupName $resourceGroup -Location $location -VM $vmConfig -Zone $zone -Verbose

    Write-Host "Deployed VM $vmName to Availability Zone $zone"
}

# Configure load balancer backend pools
$loadBalancer = Get-AzLoadBalancer -ResourceGroupName $resourceGroup -Name "avail-lb*"
foreach ($zone in $availabilityZones) {
    $backendPoolName = "Zone${zone}Pool"
    $vmName = "AutoVM-Zone$zone"
    $vm = Get-AzVM -ResourceGroupName $resourceGroup -Name $vmName
    $nic = Get-AzNetworkInterface -ResourceId $vm.NetworkProfile.NetworkInterfaces[0].Id

    # Add VM to appropriate backend pool
    $nic.IpConfigurations[0].LoadBalancerBackendAddressPools = $loadBalancer.BackendAddressPools | Where-Object {$_.Name -eq $backendPoolName}
    Set-AzNetworkInterface -NetworkInterface $nic
}

Write-Host "High availability deployment completed successfully"

Advanced Exercise 2: Availability Set Optimization

# Create optimized availability set with proximity placement groups
$resourceGroup = "Availability-Lab-RG"
$location = "East US"

# Create proximity placement group for low latency
$ppg = New-AzProximityPlacementGroup -ResourceGroupName $resourceGroup -Name "web-ppg" -Location $location -ProximityPlacementGroupType "Standard"

# Create optimized availability set
$availSet = New-AzAvailabilitySet -ResourceGroupName $resourceGroup -Name "optimized-availability-set" -Location $location -PlatformFaultDomainCount 3 -PlatformUpdateDomainCount 5 -ProximityPlacementGroupId $ppg.Id -Sku "Aligned"

# Deploy VMs with optimized placement
$vmConfigs = @()
for ($i = 1; $i -le 3; $i++) {
    $vmName = "OptimizedVM$i"
    $vmConfig = New-AzVMConfig -VMName $vmName -VMSize "Standard_D2s_v3" -AvailabilitySetId $availSet.Id -ProximityPlacementGroupId $ppg.Id
    $vmConfig = Set-AzVMOperatingSystem -VM $vmConfig -Windows -ComputerName $vmName -Credential (Get-Credential)
    $vmConfig = Set-AzVMSourceImage -VM $vmConfig -PublisherName "MicrosoftWindowsServer" -Offer "WindowsServer" -Skus "2019-Datacenter" -Version "latest"

    $vmConfigs += $vmConfig
}

# Deploy all VMs concurrently
$jobs = @()
foreach ($config in $vmConfigs) {
    $job = Start-Job -ScriptBlock {
        param($rg, $loc, $vmConfig)
        New-AzVM -ResourceGroupName $rg -Location $loc -VM $vmConfig
    } -ArgumentList $resourceGroup, $location, $config
    $jobs += $job
}

# Wait for all deployments to complete
$jobs | Wait-Job | Receive-Job

Write-Host "Optimized availability set deployment completed"

Advanced Exercise 3: Availability Monitoring and Alerting

# Set up comprehensive availability monitoring
$resourceGroup = "Availability-Lab-RG"
$workspaceName = "availability-workspace"

# Create Log Analytics workspace for centralized monitoring
$workspace = New-AzOperationalInsightsWorkspace -ResourceGroupName $resourceGroup -Name $workspaceName -Location "East US" -Sku "Standard"

# Enable VM insights for all VMs
$vms = Get-AzVM -ResourceGroupName $resourceGroup
foreach ($vm in $vms) {
    # Install monitoring agent
    Set-AzVMExtension -ResourceGroupName $resourceGroup -VMName $vm.Name -Name "MicrosoftMonitoringAgent" -Publisher "Microsoft.EnterpriseCloud.Monitoring" -ExtensionType "MicrosoftMonitoringAgent" -TypeHandlerVersion "1.0" -Settings @{"workspaceId" = $workspace.CustomerId} -ProtectedSettings @{"workspaceKey" = (Get-AzOperationalInsightsWorkspaceSharedKeys -ResourceGroupName $resourceGroup -Name $workspaceName).PrimarySharedKey}
}

# Create availability alert rules
$actionGroup = New-AzActionGroup -ResourceGroupName $resourceGroup -Name "AvailabilityAlerts" -ShortName "AvailAlert" -EmailReceiver @{Name="Operations"; EmailAddress="ops@company.com"}

# VM availability alert
$vmAvailabilityAlert = New-AzMetricAlertRuleV2 -ResourceGroupName $resourceGroup -Name "VM-Availability-Alert" -TargetResourceId "/subscriptions/$($(Get-AzContext).Subscription.Id)/resourceGroups/$resourceGroup" -WindowSize "00:05:00" -Frequency "00:01:00" -MetricName "VmAvailabilityMetric" -Operator LessThan -Threshold 0.99 -Severity 2 -ActionGroup $actionGroup

# Load balancer health alert
$lbHealthAlert = New-AzMetricAlertRuleV2 -ResourceGroupName $resourceGroup -Name "LB-Health-Alert" -TargetResourceId (Get-AzLoadBalancer -ResourceGroupName $resourceGroup).Id -WindowSize "00:05:00" -Frequency "00:01:00" -MetricName "VipAvailability" -Operator LessThan -Threshold 0.99 -Severity 1 -ActionGroup $actionGroup

Write-Host "Comprehensive availability monitoring configured successfully"

Key Takeaways

After completing this lab, you should understand:

Availability zones provide the highest level of resiliency by distributing resources across physically separate datacenters within a region, achieving 99.99% SLA
Availability sets protect against local failures within a single datacenter using fault and update domains, providing 99.95% SLA at lower cost
Load balancer configuration is critical for automatic failover and traffic distribution across available instances during failures
Health probes and monitoring enable proactive detection of issues and automatic traffic redirection to healthy instances
Proper capacity planning ensures adequate resources remain available during zone or domain failures to maintain performance
Cost optimization requires balancing availability requirements with infrastructure expenses based on business criticality and SLA targets

Availability Decision Matrix

Availability Strategy Selection

Criteria	Single VM	Availability Set	Availability Zones	Multi-Region	Recommendation
SLA requirement	99.9%	99.95%	99.99%	99.995%+	Choose based on business SLA requirements
Cost impact	Lowest	Medium	Higher	Highest	Availability sets for cost-sensitive workloads
Complexity	Simple	Moderate	Moderate	High	Start simple, evolve based on requirements
Regional resilience	None	None	High	Maximum	Zones for regional disasters, multi-region for global

Failure Protection Coverage

Failure Type	Single VM	Availability Set	Availability Zone	Multi-Region
Hardware failure	❌ No protection	✅ Protected	✅ Protected	✅ Protected
Rack failure	❌ No protection	✅ Protected	✅ Protected	✅ Protected
Datacenter failure	❌ No protection	❌ No protection	✅ Protected	✅ Protected
Regional disaster	❌ No protection	❌ No protection	❌ No protection	✅ Protected
Planned maintenance	❌ Downtime required	✅ Zero downtime	✅ Zero downtime	✅ Zero downtime

Cost vs Availability Trade-offs

Scenario	Optimization Strategy	Expected Outcome
Development/Testing	Single VMs with automated backups	70-80% cost savings with acceptable availability for non-production
Standard Production	Availability sets with load balancing	Balanced cost and availability with 99.95% SLA
Mission-Critical	Availability zones with zone-redundant services	Premium availability (99.99% SLA) with 20-30% cost increase
Global Enterprise	Multi-region deployment with traffic management	Maximum availability (99.995%+) with 2-3x cost increase