# Idle Managed Online Endpoint Deployments in Azure Machine Learning Canonical: https://www.pointfive.co/efficiency-hub/inefficiencies/idle-managed-online-endpoint-deployments-in-azure-machine-learning Managed online endpoints serve models for real-time inference. Each deployment under an endpoint runs on its own dedicated VM instances, chosen by... By: PointFive Updated: 2026-09-28 [Cloud Efficiency Hub](https://www.pointfive.co/efficiency-hub) The short version Managed online endpoints serve models for real-time inference. PointFive Research Cloud cost research at PointFive Azure service [Azure Machine Learning](https://www.pointfive.co/efficiency-hub/cloud-services/azure-machine-learning) Category [AI](https://www.pointfive.co/efficiency-hub/service-category/ai) Reference CER-0458 Type Idle or Unused Resource ## Explanation Why the waste happens and who it affects. Each deployment under an endpoint runs on its own dedicated VM instances, chosen by instance type and instance count, and Microsoft states that costs apply to the virtual machines assigned to the deployment. Those VMs bill for as long as the deployment exists, whether it receives heavy traffic, a trickle of requests, or none at all. Deployments accumulate quietly. Blue/green rollouts leave the previous deployment in place with 0 percent of traffic, test and demo endpoints outlive the project, deployments for retired model versions are never removed, and failed deployments can still incur charges if they got as far as creating compute. Endpoints serving GPU models are the most expensive case, because each idle instance is a GPU VM. ## Billing model The pricing dimensions that drive this cost. Managed online endpoints are billed for the compute and networking they use, with no Azure Machine Learning surcharge. Deployment instances Each deployment is billed for its VM instance type multiplied by its instance count for as long as the deployment exists, including deployments that receive 0 percent of traffic Quota reservation For many VM SKUs an extra 20 percent of quota is reserved for upgrades; it does not incur cost unless system operations use it Failed deployments A failed deployment that passed the compute creation stage incurs charges until it is deleted Managed virtual network If outbound traffic is secured with a managed virtual network, Private Link and FQDN outbound rules are charged separately ## How to detect 5 checks to find it in your estate. - List endpoints and deployments (az ml online-endpoint list and az ml online-deployment list, or studio \> Endpoints) and review each endpoint's traffic allocation; deployments with 0 percent traffic and no mirrored traffic are candidates for deletion - Chart the endpoint RequestsPerMinute metric split by the deployment dimension over 14 to 30 days to find deployments with little or no traffic - Check the deployment metrics DeploymentCapacity and CpuUtilizationPercentage or GpuUtilizationPercentage for instances that are provisioned but nearly idle - Find deployments in a failed provisioning state that still exist - In Cost Analysis, filter to the workspace resource and use the azuremlendpoint and azuremldeployment tags to see the cost of each endpoint and deployment ## How to fix 5 ways to remove the waste. - Delete deployments that no longer receive traffic, such as the old side of a completed blue/green rollout, and delete endpoints used only for tests or demos - Delete failed deployments once debugging is finished - For deployments that must stay, lower the instance count and configure Azure Monitor autoscale with metric rules and schedule-based profiles (for example fewer instances on weekends), weighing lower minimums against availability; Microsoft recommends at least 3 instances for high availability - Move latency-tolerant or periodic scoring to batch endpoints, which run as jobs on compute clusters that deallocate when the job completes and can scale to zero - Right-size the instance type using utilization metrics, and move CPU-sufficient models off GPU SKUs ## Documentation Vendor references for pricing and configuration. - [Online endpoints for real-time inference - Azure Machine Learning learn.microsoft.com](https://learn.microsoft.com/en-us/azure/machine-learning/concept-endpoints-online) - [Manage and optimize costs - Azure Machine Learning learn.microsoft.com](https://learn.microsoft.com/en-us/azure/machine-learning/how-to-manage-optimize-cost) - [View costs for managed online endpoints - Azure Machine Learning learn.microsoft.com](https://learn.microsoft.com/en-us/azure/machine-learning/how-to-view-online-endpoints-costs) - [Autoscale online endpoints - Azure Machine Learning learn.microsoft.com](https://learn.microsoft.com/en-us/azure/machine-learning/how-to-autoscale-endpoints) - [Azure Machine Learning monitoring data reference learn.microsoft.com](https://learn.microsoft.com/en-us/azure/machine-learning/monitor-azure-machine-learning-reference) - [What are batch endpoints? - Azure Machine Learning learn.microsoft.com](https://learn.microsoft.com/en-us/azure/machine-learning/concept-endpoints-batch) ## Related inefficiencies [Browse the library](https://www.pointfive.co/efficiency-hub) - Azure Machine Learning CER-0456 ### [Azure Machine Learning Compute Instances Without Idle Shutdown](https://www.pointfive.co/efficiency-hub/inefficiencies/azure-machine-learning-compute-instances-without-idle-shutdown) An Azure Machine Learning compute instance is a single-user development VM for notebooks, VS Code and experiments. Microsoft notes that when you create a compute instance the VM stays on so it's available for your work, and it is billed... AI - Azure Machine Learning CER-0457 ### [Azure Machine Learning Compute Clusters With Nonzero Minimum Nodes](https://www.pointfive.co/efficiency-hub/inefficiencies/azure-machine-learning-compute-clusters-with-nonzero-minimum-nodes) Azure Machine Learning compute clusters (AmlCompute) scale up when training or batch inference jobs are submitted and scale back down to the configured minimum node count when jobs finish. When the minimum is set above zero, usually to... AI - Azure Cognitive Services CER-0238 ### [Always-On PTUs for Seasonal or Cyclical Azure OpenAI Workloads](https://www.pointfive.co/efficiency-hub/inefficiencies/always-on-ptus-for-seasonal-or-cyclical-azure-openai-workloads-ccb97) Many Azure OpenAI workloads - such as reporting pipelines, marketing workflows, batch inference jobs, or time-bound customer interactions-only run during specific periods. When PTUs remain fully provisioned 24/7, organizations incur... AI --- Source: the public page above. Product screenshots and illustrative interfaces are examples, not live customer data.