Share via

Allocation failed for insufficient capacity in the region even though I just purchased a reservation for this size.

Josh ‎ 21 Reputation points
2026-04-14T18:32:41.0766667+00:00

I stopped the VM yesterday. Today I tried to start it and it didn't have capacity. Ok so I purchased a reservation for that size. But still it says the same thing, that it doesn't have capacity. CoPilot told me to open a support request, but I have to pay $29/mo for technical support apparently. Anyone know the fix for this?

Thank you

Community Center | Not monitored
0 comments
Answer accepted by question author
  1. Jose Benjamin Solis Nolasco 7,996 Reputation points Volunteer Moderator
    2026-04-14T18:56:13.7666667+00:00

    Hello Josh, Welcome to Microsoft Q&A,

    What you are running into is the difference between a billing reservation and a capacity reservation. The standard reservation you purchased is strictly a financial instrument. It pre-pays for the compute at a steep discount, but it does not actually reserve physical hardware on the datacenter floor. When you stopped and deallocated the VM yesterday, it released the hardware back into the public pool. Today, that region is experiencing a spike in demand, and there are literally no physical slots left for that specific VM size.

    The quickest way to get your server back online is to just go into the VM settings and resize it to a different SKU. For instance, if you are using a D2s_v3, try switching it to a D2s_v4 or v5. Newer hardware generations almost always have more available capacity. Just be aware of how that impacts the billing reservation you just bought! Otherwise, if you don't want to change the size, your best bet is to just wait a few hours and try starting it again once someone else turns their server off and frees up some space.

    😊 If my answer helped you resolve your issue, please consider marking it as the correct answer. This helps others in the community find solutions more easily. Thanks!

    1 person found this answer helpful.
    0 comments

1 additional answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Q&A Assist
    2026-04-14T18:32:56.5866667+00:00
    AI generated content. This question contains an answer created with AI from Q&A Assist. Learn more

    A reservation doesn’t guarantee that an individual VM start or resize will succeed if the cluster or zone where that VM is pinned has no free capacity. It only gives a billing benefit and, in some cases, capacity priority at the region level, not a hard guarantee for a specific stopped VM on its original cluster.

    From the documented behavior, the failure is caused by lack of capacity in the region/cluster for that exact size and constraints (zone, availability set/PPG, etc.), even after purchase.

    Use the supported workarounds for allocation failures:

    1. If this is a standalone VM (no availability set/PPG):
      • Retry starting the VM later; capacity can free up.
      • If possible, resize the VM to a different size that has capacity in the same region/zone.
      • If allowed, move the workload to a different region or zone where capacity is available.
    2. If the VM is in an availability set:
      • Stop (deallocate) all VMs in the availability set.
      • After all show Stopped (deallocated), start them again together so Azure can place them on a different cluster with capacity:
        1. In the portal, go to Virtual machines.
        2. Add a filter for the availability set.
        3. Select all VMs in that set and choose Stop; wait until all are Stopped (deallocated).
        4. Select them again and choose Start.
    3. If the VM is part of a scale set, follow the equivalent guidance for VMSS (stop/deallocate all instances, then restart/resize so the set can move to a cluster with capacity).
    4. To avoid this in the future, use On-demand Capacity Reservations (different from just a billing reservation) so capacity is reserved ahead of time for that size in the region. Note that even capacity reservations can fail to create if the region/zone is already out of capacity for that size; they must succeed first to be effective.

    If none of the above options (resize, different region/zone, deallocate-all-and-restart) are acceptable, the only remaining path is to wait for capacity to free up or work with Azure support, because there is no portal-side “fix” that forces capacity into an exhausted cluster.

    References:

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.