Virtual Machine Scale Sets

Azure Virtual Machine Scale Sets (VMSS) let you create and manage a group of identical, load-balanced VMs. They automatically increase or decrease the number of VM instances in response to demand or a defined schedule, providing high availability and application resiliency.

---

What is a Virtual Machine Scale Set?

A VMSS is a collection of VMs that are:

• Identical — all created from the same base image and configuration
• Load-balanced — traffic is distributed evenly across instances
• Auto-scaling — instances are added or removed based on demand
• Zone-aware — can span Availability Zones for high availability

Key Benefits

Benefit	Description
Automatic scaling	Scale from 0 to 1,000 instances based on metrics or schedules
High availability	Distribute across fault domains and Availability Zones
Consistency	All VMs are identical — same image, size, and configuration
Integration	Works with Azure Load Balancer and Application Gateway
Cost efficiency	Scale in during low demand to reduce costs

---

Orchestration Modes

VMSS supports two orchestration modes:

Uniform Orchestration

• All VMs are created from the same model (image, size, configuration)
• Best for large-scale stateless workloads (web frontends, batch processing)
• Azure manages instance naming, networking, and distribution
• Supports up to 1,000 instances with standard marketplace images, or 600 with custom images

Flexible Orchestration

• Mix different VM sizes and configurations within the same scale set
• Provides the benefits of scale sets with more control over individual VMs
• Can add existing VMs to the scale set
• Supports up to 1,000 instances
• Recommended for most new deployments

# Create a Flexible VMSS
az vmss create \
  --resource-group rg-vmss-demo \
  --name myScaleSet \
  --orchestration-mode Flexible \
  --image Ubuntu2204 \
  --vm-sku Standard_D2s_v5 \
  --instance-count 3 \
  --zones 1 2 3 \
  --admin-username azureuser \
  --generate-ssh-keys \
  --lb myLoadBalancer

---

Autoscaling

Autoscaling is the core capability of VMSS. It adjusts the number of instances to match demand.

Metric-Based Autoscaling

Scale based on performance metrics:

Metric	Scale Out When	Scale In When
CPU utilisation	Average > 70% for 5 minutes	Average < 30% for 10 minutes
Memory utilisation	Average > 80% for 5 minutes	Average < 40% for 10 minutes
Network In/Out	Bytes exceed threshold	Bytes below threshold
Disk Queue Length	Queue consistently growing	Queue consistently empty
Custom metrics	Application-specific thresholds	Application-specific thresholds

Schedule-Based Autoscaling

Scale based on known patterns:

Schedule	Instance Count
Weekdays 8:00–18:00	10 instances
Weekdays 18:00–8:00	3 instances
Weekends	2 instances

Autoscale Rules

# Create an autoscale profile
az monitor autoscale create \
  --resource-group rg-vmss-demo \
  --resource myScaleSet \
  --resource-type Microsoft.Compute/virtualMachineScaleSets \
  --name autoscale-profile \
  --min-count 2 \
  --max-count 20 \
  --count 3

# Add a scale-out rule (CPU > 70%)
az monitor autoscale rule create \
  --resource-group rg-vmss-demo \
  --autoscale-name autoscale-profile \
  --condition "Percentage CPU > 70 avg 5m" \
  --scale out 2

# Add a scale-in rule (CPU < 30%)
az monitor autoscale rule create \
  --resource-group rg-vmss-demo \
  --autoscale-name autoscale-profile \
  --condition "Percentage CPU < 30 avg 10m" \
  --scale in 1

Scaling Best Practices