Request More Azure vCPU Quota Without Leaving Atlas AVD

Running out of Azure vCPU quota stops a scale-out or an image rollout. Adaptive Quota in Atlas AVD shows the vCores your host pools need against each quota, lists what's spare in every region, suggests a new limit from your usage trend, and submits and tracks the increase with Azure in a few clicks.

By David,

It's 8:40 on a Monday. Your scaling plan asks for six more session hosts, Azure answers with a quota error, and the users who sign in next land on hosts that are already full. The fix is a bigger vCPU quota, but asking for one usually means leaving your AVD tooling, finding the right subscription, region and VM family in the Azure portal, guessing a number and checking back later.

In Atlas AVD it's four steps on one screen. This post walks through them with mockups of each screen.

The short version: open Adaptive Quota, find the VM family that's running short on Current Demand, check Browse Available for spare quota in other regions, choose Request Quota Increase, accept or adjust the suggested limit, and follow the request in Request History until Azure approves it. Atlas files the request with Azure for you; Azure still decides.

Azure quota behavior described here is taken from Microsoft's documentation as of October 2026. Atlas AVD screens are shown as simplified mockups with illustrative sample data.

1. Why vCPU quota catches AVD teams out

Azure limits how many vCPUs a subscription can run in each region, in two tiers: total regional vCPUs and vCPUs per VM family (for example, the DASv5 family that Standard_D8as_v5 belongs to). A deployment that would go over either limit fails (vCPU quotas).

For AVD, quota runs out when the environment changes, not on a quiet day:

  • Scale-out at peak, when the autoscaler builds hosts for the morning logon wave.
  • Image rollouts, when replacement hosts briefly run next to the hosts they replace.
  • New host pools, a new VM size or a GPU pool for a design team.
  • Growth, when a department joins and nobody told the quota.

You can raise a VM-family quota yourself in the portal under Quotas (Increase VM-family vCPU quotas). Many requests are approved within minutes; others go to a support request. The hard part is rarely the form. It's knowing which family, in which region, needs how much, before it fails. Our Azure quota guide covers the portal route step by step.

2. Meet Adaptive Quota

Adaptive Quota™ is the quota page in Atlas AVD. It sits under Atlas Compass in the navigation, and its Overview tab shows vCPU in use against your quota limit over the last 24 hours, how often provisioning fell back to another VM size, and quota usage for each host pool.

The other three tabs, Current Demand, Browse Available and Request History, are where you request more.

3. Step 1: See your demand

Current Demand lists every VM size your session hosts use, by region, with the numbers that matter for quota:

  • vCores Req: the cores your VMs use (VM count × cores per VM).
  • Limit: Azure's quota for that VM family in that region.
  • % Util and Headroom: how much of the quota is used, and how many vCores are left before the limit.
  • Risk: Low below 60%, Medium from 60% and High from 80%. You can change both thresholds in settings.

The Constrained only toggle hides everything below your risk threshold, so in a large estate you see only the quotas that need action. Anything at risk also appears under Needs attention with its vCores left and a Request increase button. Here, DASv5 in East US 2 has 24 vCores left: three more 8-core hosts.

4. Step 2: Check what's available

Before you ask for more, check whether you already have it somewhere else. Browse Available shows every VM family's Current Limit, Used and Available vCores in the region you pick. Turn on AVD recommended only to hide sizes that don't suit session hosts.

This tab is also how you plan a new environment: pick the region you're deploying into and see what fits before you build anything. If another region has spare quota for the same family, the request panel's Shift Workload option lists it, so you can place a new pool there instead of waiting on an increase.

5. Step 3: Request the increase

Choose Request Quota Increase (or Request increase under Needs attention, or Submit Increase Request from a row's menu on Browse Available). The panel opens with the VM family, region, current limit and usage already filled in.

Then pick the new limit:

  • Use the suggestion. Atlas recommends a limit from your current footprint plus a 25% buffer (Based on 22 VMs × 8 cores + 25% buffer), and once it has enough usage history, a forecast such as Forecast: quota exhausts in ~12 days at current growth.
  • Use Quick Increase. One click for +25%, +50% or +100% of the current limit.
  • Type your own in Requested New Limit (cores).

Add a Reason (Capacity Growth, Migration, New Deployment or Seasonal Demand) and an optional Justification, then choose Submit Request. Atlas sends the request to Azure through the Azure Quota API, the same request the portal makes, and confirms with Quota Request Submitted. You don't switch to the portal or open a ticket.

Tip: ask for room for your next image rollout as well as your growth. Rollouts run replacement hosts next to the old ones for a while, so the quota you need at peak is higher than the quota you need on a normal day.

6. Step 4: Track it in Request History

Every request goes to Request History with its Current Limit, Requested Limit and Status. Atlas checks each pending request with Azure every few minutes and updates the status when Azure answers, so you don't have to keep the portal open.

  • In Progress means Azure is still working on it. Large increases can go to Microsoft for review, and Atlas keeps tracking those for up to seven days.
  • Approved means the new limit is live, and Current Demand shows the extra headroom.
  • Failed shows why. Retry Request resubmits it, and Create Support Bundle downloads the request details as a file to attach to a Microsoft support case.

Atlas allows one pending request for each VM family, region and subscription, so two admins can't file the same increase twice.

7. What makes it adaptive

The four steps work on their own. Turn on the Adaptive Quota setting and Atlas also watches quota for you:

  • Forecasts. Each time Adaptive Quota refreshes demand it records a usage snapshot for each quota (at most one an hour), and fits a trend to project when that quota will run out. The forecast looks 30 days ahead by default (Quota Forecast Horizon), and Atlas keeps 90 days of snapshots by default (Quota Snapshot Retention).
  • Alerts. Quota Alerts warn you when a quota crosses your High threshold or is forecast to run out, at most once a day per quota. It's critical at 90% or when exhaustion is less than a week away.
  • Auto-submit, if you want it. With Auto-Submit Quota Requests on, Atlas files the suggested increase itself when a quota is forecast to run out within the horizon and the forecast has enough data behind it. It skips any quota that already has a pending request. It's off by default.
  • Fallback during provisioning. When a deployment would go over quota, Adaptive Quota picks a compatible VM size that has quota instead of letting the deployment fail, and the Overview counts each fallback.
  • Answers in Atlas Advisor. Ask Atlas Advisor whether you have room for ten more 8-core hosts in East US 2 and it checks the live headroom. If you don't, it points you to Adaptive Quota.

8. Setup: one Azure role

Atlas files quota requests as its own app registration, so that app needs permission to read and change quota on each subscription:

  • Assign the Quota Request Operator role to the Atlas app registration at subscription scope (Azure quotas overview).
  • Make sure the Microsoft.Quota resource provider is registered on the subscription.

If either is missing, Atlas tells you which one when you submit, instead of failing quietly. Quota requests work the same in Azure Government and GCC High: Atlas sends them to your cloud's own Azure endpoints and lists government regions such as USGov Virginia.

Adaptive Quota is included in the Enterprise and MSP plans; see pricing for what each plan includes.

What Adaptive Quota doesn't do

  • It doesn't approve quota. Azure decides every request. Atlas files it, tracks it and shows you the answer.
  • It doesn't guarantee capacity. Quota is permission to deploy, not reserved hardware. Azure can still be short of a size in a region even when your quota allows it.
  • It doesn't file requests on its own unless you turn on Auto-Submit Quota Requests.
  • It doesn't replace your Microsoft support plan. A denied or long-running request still goes to Microsoft, and the support bundle gives them the details.

FAQ

How do I request more vCPU quota for Azure Virtual Desktop?

In the Azure portal, open Quotas, choose Compute, find the VM family and region, and request a new limit. In Atlas AVD, open Adaptive Quota, choose Request Quota Increase on the VM family that's running short, pick a limit and submit. Atlas sends the same request to Azure and tracks it in Request History.

How long does an Azure vCPU quota increase take?

Many VM-family increases are approved automatically within minutes. Larger increases, GPU families and some regions can go to Microsoft for review, which can take days. Atlas shows the status as In Progress until Azure answers.

How much quota should I ask for?

Enough for your peak plus your next change. Atlas suggests your current footprint plus a 25% buffer, and shows when the quota is forecast to run out at your current growth. Add room for image rollouts, which briefly run old and new hosts side by side.

Does Atlas AVD increase quota automatically?

Only if you turn on Auto-Submit Quota Requests. It's off by default. When it's on, Atlas files the suggested increase when a quota is forecast to run out, and never while a request for that quota is still pending.

What permissions does Atlas need to request quota?

The Quota Request Operator role on each subscription, assigned to the Atlas app registration, and the Microsoft.Quota resource provider registered on the subscription.

Does this work in GCC High?

Yes. Atlas sends quota requests to your Azure Government endpoints, and the regions you pick from are government regions.

Related reading: Azure Virtual Desktop fleet management covers why quota belongs in day-to-day operations, the AVD sizing guide helps you estimate the cores you'll need, and converting to v7 sizes explains why a new VM family needs its own quota.

Sources: Microsoft Learn, vCPU quotas for Azure virtual machines, Increase VM-family vCPU quotas and Azure quotas overview, as of October 2026.