What’s new — Managed Prometheus now supports VMs & VMSS
Today we are excited to announce the private preview of Azure Managed Prometheus support for virtual machines (VM) and virtual machine scale sets (VMSS). Until now, Managed Prometheus on Azure was primarily targeted at containerized workloads — e.g. Kubernetes (AKS) or Azure Arc–enabled clusters. With this preview, you can now extend Prometheus-style monitoring to your IaaS workloads running on VMs/VMSS, giving you unified, scalable, resilient metric collection and observability across both containers and traditional compute —including full support for GPU and InfiniBand (IB) metric collection for HPC scenarios.
Behind the scenes, Azure Monitor provides the storage, ingestion pipeline, and query engine, while surfacing a fully compatible Prometheus experience — including scraping, PromQL, alerting rules, and dashboards.
Why this matters — especially for HPC workloads
Azure HPC customers running large fleets of GPU-accelerated VMs and VMSS nodes can now:
- Collect node-level metrics (CPU, memory, disk, frontend NIC, InfiniBand) and GPU metrics (utilization, memory, clocks, ECC, throttling) through standard Prometheus exporters
- Store all Prometheus metrics in an Azure Monitor Workspace
- Visualize cluster performance using Azure Managed Grafana with out of the box dashboards that include cluster-level views, node-level views, and data links to easily move between them.
- Run PromQL queries directly against Azure Monitor
- Monitor mixed fleets (AKS + VMSS + standalone VMs) in one unified system
All of this is achieved through a fully managed Prometheus backend, with no servers, scaling, or storage to manage.
Access Requirement
This feature is currently in private preview, and your Azure subscription must be allowlisted before you can use Azure Managed Prometheus for VMs/VMSS.
Request access to the private preview
Once approved, you will be notified and can proceed with the onboarding steps in the GitHub repository.
Try it yourself
We invite you to try it out and share your feedback with us. To get started, follow the step-by-step guide in our GitHub repository to help you onboard to the preview quickly.
Once you’ve onboarded, you can begin scraping node and GPU metrics, run sample PromQL queries, and import ready-made HPC dashboards into Azure Managed Grafana.
We hope you enjoy using Azure Managed Prometheus for VM/VMSS and find the new capabilities valuable for your AI and HPC workloads. As this is a private preview, your feedback is especially important. Please share input by opening an issue in the GitHub repository.
Microsoft