Public Preview - Azure Compute Fleet

Azure is pleased to announce the public preview of Azure Compute Fleet, anew service that streamlines the provisioning and management of Azure compute capacity across different virtual machine (VM) types, availability zones, and pricing models to achieve desired scale, performance, and cost.

Azure Compute Fleet provides customers with many features to deploy and manage diverse groups of VMs at scale.  Key capabilities include:

• Integration of multiple pricing models within a single fleet request.
• Automated configuration of a fleet of VMs to achieve the best mix of VM sizes, pricing models, and performance, based on current capacity availability.
• Adjustable settings to prioritize either deployment speed, operational cost, or a balance of both.
• Capability to manage and deploy up to 10,000 VMs in a region within a single fleet.
• Numerous options that allow users to automatically and programmatically control how fleets respond to changing variables, such as cost overruns, capacity shortages for specific VM sizes, or the eviction of Spot VMs.

These enhancements not only simplify obtaining compute capacity according to workload demands, but also reduce the time and financial costs of managing VM fleets as needs evolve.  In addition, they increase the flexibility of deployment options based on criteria such as VM types, prices, replacements, and allocation strategies, while increasing fleet reliability by maintaining capacity through automated spot VMs, VM mix, and cross-zonal deployment features.  Azure Compute Fleet thus makes it significantly easier for Azure compute customers to scale to meet compute demand, optimize their cloud costs, and customize their fleet deployments.

Azure Compute Fleet will be available in preview in the following regions: West US, WestUS2, East US, and East US2.

 You can learn more about Azure Compute Fleet in our preview blog and documentation.