Azure Virtual Machine Scale Sets (VMSS) let you create and manage a group of load-balanced, automatically scaling VM instances.
What are Azure Virtual Machine Scale Sets?
Virtual Machine Scale Sets are Azure’s solution for managing large groups of VMs that scale automatically based on load, schedules, or manual configuration. Instances are created from a shared configuration model, allowing hundreds of VMs to be managed without extra configuration effort. Built-in load balancing via Azure Load Balancer (Layer 4) or Application Gateway (Layer 7, including TLS termination) automatically distributes traffic across instances.
There are two orchestration modes: Flexible (recommended by Microsoft for new deployments) manages standard Azure VMs with full access to standard VM APIs, supports mixed VM sizes and operating systems, allows combining Spot and on-demand instances, and provides high-availability guarantees for up to 1,000 instances. Uniform (classic) manages instances as a group via scale-set-specific VM APIs, currently supports higher maximum instance counts (up to 3,000 with fault domain guarantees), but offers less flexibility and is not compatible with standard VM features such as Azure Backup or Azure Site Recovery. The orchestration mode is set at creation time and cannot be changed afterward.
For high availability, scale set instances can be spread across fault domains within a datacenter or across multiple availability zones to protect against entire datacenter failures.
Core Features
- Automatic scaling based on metrics (CPU, memory, custom metrics), schedules, or manual configuration
- Two orchestration modes: Flexible (recommended, standard VM API, mixed VM types) and Uniform (classic, high instance counts)
- Built-in load balancing via Azure Load Balancer or Application Gateway
- Distribution across fault domains and availability zones for high availability
- Support for Spot VMs to optimize costs for fault-tolerant workloads
- Automated update policies (Automatic, Rolling, Manual)
Typical Use Cases
Auto-Scaling for Web Applications
Web front ends scale automatically during traffic spikes and reduce instance count during low utilization, saving costs during quiet periods.
Worker Nodes for Kubernetes/AKS
VMSS serves as the foundation for node pools in Azure Kubernetes Service. Flexible orchestration supports mixed VM sizes for optimized workload placement.
Batch Processing and Big Data
Compute resources for batch jobs scale up only when needed. Spot VMs in scale sets enable significant cost savings for interruptible workloads.
Stateless Microservices and CI/CD Build Agents
Stateless applications and build agents benefit from fast horizontal scaling and simple management of identically configured instances.
Benefits
- No additional cost for the scale set service itself, only for resources actually used
- Unified management of large VM fleets without manual configuration overhead
- Combinable with Spot instances for significant cost savings
- Seamless integration with autoscale, load balancing, and availability zones
Frequently Asked Questions about Virtual Machine Scale Sets
What is the difference between Uniform and Flexible orchestration?
Flexible orchestration manages standard Azure VMs with full access to standard VM APIs, supports mixed VM sizes/operating systems and a mix of Spot and on-demand instances, and provides high-availability guarantees for up to 1,000 instances. Uniform orchestration manages instances via dedicated scale-set VM APIs, supports higher maximum instance counts, but offers less flexibility and no compatibility with services such as Azure Backup. Microsoft recommends Flexible orchestration for new deployments.
What does a Virtual Machine Scale Set cost?
There is no additional charge for the scale set service itself. You are billed only for the compute, storage, and network resources used, as with regular VMs, plus a separately billed load balancer if one is used.
Can I use Spot VMs in scale sets?
Yes, Azure Spot VMs can be used in scale sets for fault-tolerant, interruptible workloads and enable significant cost savings compared to regular on-demand instances. A configurable eviction policy controls behavior when capacity is constrained (deallocate or delete).
Does VMSS support stateful applications?
Flexible orchestration now supports stateful applications and quorum-based workloads better than before, including support for Azure Backup and Azure Site Recovery. For heavily stateful workloads with complex orchestration requirements, Azure Kubernetes Service with StatefulSets is often still the more suitable choice.
What is the difference from Availability Sets?
Availability Sets provide a fixed number of VMs with manual scaling and no built-in autoscale capability. Virtual Machine Scale Sets provide automatic scaling, distribution across availability zones, and more flexibility in instance management.
Integration with innFactory
As a Microsoft Solutions Partner, innFactory supports you in designing and implementing auto-scaling architectures with Virtual Machine Scale Sets. We help with selecting the right orchestration mode, configuring autoscale rules, integrating with load balancers, and migrating from Availability Sets to VMSS.
Contact us for a non-binding consultation on Azure Virtual Machine Scale Sets and Microsoft Azure.
Typical Use Cases
Technical Specifications
Note: All product information on this page has been compiled with care, but is provided without guarantee and may be outdated or incomplete. Cloud services evolve rapidly — features, pricing, SLAs, and availability change frequently. Authoritative and up-to-date information can only be found on the official product page of Azure (official documentation). This page does not represent an offer by Azure.
