Azure Site Recovery Test Failover fails in UAE North due to VM SKU capacity restrictions

Hi Help Desk & thx for join me at Q&A platform,

seems an Azure Compute capacity issue, not an ASR replication problem. The replication can stay Healthy while Test Failover still fails later when ASR asks Compute to allocate the temporary VM. The error you’re getting is the important part: quota being available doesn’t guarantee physical capacity for a specific SKU. Microsoft explicitly notes that quota is only an approved limit and doesn’t reserve capacity, so a deployment can still fail with SkuNotAvailable or a capacity restriction.

For ur requirement, Standard_E4s_v5 and Standard_E4as_v5 are valid 4-vCPU / 32-GB sizes. The fact that several E4 variants all fail in UAE North strongly suggests that the needed hardware capacity isn’t currently allocatable for your subscription in that region.

There isn’t a public API that gives guaranteed real-time allocatable capacity before deployment. az vm list-skus --location uaenorth --all is useful for checking whether a SKU is offered and whether Azure reports a subscription restriction, but it does not guarantee that capacity will still be available when ASR performs the failover. Microsoft documents az vm list-skus for regional/SKU availability checks, not as a capacity reservation mechanism.

I’d try a newer compatible memory-optimized family if ASR allows you to select one, but if all suitable 4-vCPU/32-GB SKUs in UAE North return the same capacity restriction, there isn’t an ASR-side fix. At that point, either test failover into another supported region or open an Azure Compute/ASR support case and ask them to check current UAE North allocation availability for the target VM families.

rgds,

Alex

Hi @Help Desk

Based on the error and the checks you've already completed, this doesn't appear to be an Azure Site Recovery replication problem. The failure is occurring at the VM allocation stage when ASR attempts to create the temporary test-failover VM in UAE North.

The message "The requested VM size is currently not available in location UAE North due to Capacity Restrictions" indicates an Azure compute capacity/allocation restriction, which is different from your subscription's vCPU quota. Sufficient quota means your subscription can deploy those vCPUs; it doesn't guarantee Azure has physical capacity for a particular VM SKU in that region.

Insufficient capacity/quota and VM-size constraints may prevent Site Recovery from creating the failed-over Azure VM.

Since you've already tried several E-series sizes, I wouldn't keep selecting closely related E*v5 SKUs blindly. First, check which VM sizes Azure reports as available for your subscription in UAE North:

az vm list-skus \
  --location uaenorth \
  --resource-type virtualMachines \
  --all \
  --output table

Pay particular attention to the Restrictions column. You can narrow the results, for example:

az vm list-skus \
  --location uaenorth \
  --resource-type virtualMachines \
  --all \
  --query "[?capabilities[?name=='vCPUs' && value=='4']]" \
  --output table

However, there is an important limitation: SKU discovery and the Restrictions information should not be treated as a real-time guarantee of physical capacity. A SKU can be listed as available and still fail allocation if capacity changes at deployment time.

For the ASR test itself, you can change the recovery VM size under the replicated item's Compute and Network settings. Site Recovery supports modifying the target VM size before failover.

Since this is a Test Failover, you could temporarily choose another compatible VM family with roughly the required CPU/memory profile if one is available. The objective of the drill is to validate that the workload can be recovered; the test VM doesn't necessarily have to use the exact same SKU you intend to use for production recovery.

Also, run a simple control test: create an ordinary Azure VM in UAE North using the same SKU outside ASR. If that deployment produces the same capacity restriction, you've effectively confirmed that ASR isn't the source of the failure.

Azure Site Recovery performs prerequisite checks and then creates an Azure VM from the selected recovery point during a test failover, so an Azure Compute allocation failure at that stage can prevent an otherwise healthy replication from completing the drill.

If none of the suitable 4-vCPU/~32-GB SKUs can currently be allocated in UAE North, open an Azure Support request under Virtual Machine/Compute capacity and provide:

• Region: UAE North
• Required profile: approximately 4 vCPU / 32 GB RAM
• SKUs already tested: Standard_E4s_v5, E4as_v5, E4bs_v5, E4ds_v5, E4d_v5
• Exact allocation error
• ASR Test Failover job ID and correlation/activity ID
• UTC timestamps of the failed deployments

Microsoft Support can then confirm whether you're encountering a current regional/cluster capacity constraint and advise which suitable VM family has deployable capacity.

One other option, if your DR design allows it, is to evaluate a different Azure recovery region. Azure Site Recovery supports replication between Azure regions, but changing the recovery-region design should be treated as an architecture decision rather than simply a workaround for a temporary UAE North capacity shortage.

For the real-time capacity part of your question: no customer-facing command can guarantee that a particular SKU will allocate successfully at the exact moment ASR performs the failover. az vm list-skus is useful for SKU/restriction discovery, but an actual allocation attempt or confirmation from Azure Support is ultimately needed for a current capacity constraint.

References:

Run an Azure Site Recovery test failover

Troubleshoot Azure Site Recovery failover failures

Azure Site Recovery network mapping and IP addressing

Help make this community better for everyone: If this answer helped or resolved your issue, please accept it or upvote it. If not, share more details in a comment so we can continue the discussion and find the right solution. Thank you.