Content

Scenario: AI model inference. Features: EAS allows services deployed in a dedicated resource group to be automatically scaled without being limited by the node resources of the resource group. If resources in the dedicated resource group are insufficient during a service scale-out, the service instances added to the service are created in the public resource group and billed based on the rules of the public resource group. During a service scale-in, service instances that reside in the public resource group are released first. You can use this feature with the auto scaling feature by enabling auto scaling based on the fluctuation of metrics such as QPS or CPU utilization for a service. This way, services can be automatically scaled out without being limited by the node resources of the resource group. In this case, your services are billed by using both the subscription and pay-as-you-go billing methods, which helps you reduce costs.