Microsoft recently unveiled Azure Compute Fleet at its annual Build conference. This new service aims to simplify the provisioning and management of Azure compute capacity, allowing users to access fast computing capacity in specific regions by launching a combination of Standard and Spot virtual machines at the best price and greater availability. Different pricing models, including Reserved Instances, Savings Plan, Spot, and Pay-As-You-Go options, can help achieve better price-performance ratios.
With Azure Compute Fleet, users can deploy up to 10,000 virtual machines with a single API, utilizing both spot and standard virtual machine types together. This allows for flexibility in workload requirements and the ability to maintain target capacity of specific VMs in the fleet. Deploying Compute Fleet with a predefined fleet allocation strategy can help optimize the mix of virtual machines to achieve capacity and profitability or a balance of both.
Similar services are offered by AWS and Google Cloud Platform, such as AWS Systems Manager Fleet Manager, Amazon EC2 Fleet, and Google Cloud Deployment Manager. These services provide centralized interfaces for managing servers and automating the creation and management of cloud resources.
Azure Compute Fleet is currently available in US West, US West 2, US East, and US East 2 regions. More details on the service are available on the documentation pages. This new offering from Microsoft aims to provide users with enhanced control, scalability, and cost-efficiency in managing Azure compute capacity.
Article Source
https://www.infoq.com/news/2024/05/azure-compute-fleet-management/