Hello all, this one isn't clear to me
Capacity priority prioritizes data center capacity for your deployments.
When I stop a Vm and deallocate and then I start Vm same sku type and size what will happen it could be started in different Dc but same zone or the zone could be changed? In case it is different zone stopped Vm is usually connected to one zone when it is configured. What will happen then? I'm looking for scenario where I have test server in pairex dr region and in case of Dr I would like to stop test and start prod with same zone location and server type and size
Thanks
⚠Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
Thanks for the question! We are investigating and will update you shortly.
Adding @Drewm3
When you reserve VM instances in Azure there are two options to pick from
1) Size Flexibility
2) Capacity Priority
Choosing to use Size Flexibility means that the reservation you paid for can apply to multiple sizes of the same VM series. This means that if you have a reservation for say D3 VMs and have size flexibility selected and want to deploy a VM then we go and check to ensure that size is available and if it is not, then we will allocate your VM to a different size in the D family and apply the reserved discount to that size. This means that even if there was no D3 capacity we would still ensure you are able to deploy a VM of the same D series size and your discount will be applied. Size flexibility is good for when you might not require a specific size but know for sure you want to be able to deploy in that region when needed.
Capacity Priority is used for when you want that specific size. So if you paid for a reserved instance of D3 VM and you pick Capacity Priority you are saying I ONLY want that size VM. By choosing to go with Capacity Priority over Size Flexibly we will ensure there is enough of that specific series in the region you reserved for you to deploy that VM in that size.
As mentioned in the doc, when you have a reservation that is limited to single scope (VM Size) you can opt to have the optimization settings set to Capacity rather than flexibility.
Does that help clarify?
Uff it is more and more difficult for me.
When I order reserved capacity with capacity priority. It will be only applied when I order the specific VM?
When I order this I understand that there will be more Vms with this size in this region. Will this means that when I order for example D4sv3 with capacity priority there will be caunted with one more D4sv3? Will be that connected for my subscription? Or it is only to increase chance that when I order D4sv3 that I will get it?
Thanks
If you pay for a reserved instance and use Capacity Priority we will reserve an instance of the selected size for you in that region. That way if you want to spin up an instance you can be confident that there will be that specific size available for your subscription.
Otherwise, flexibility capacity will ensure you that size or a similar size available but the discount for reserving the capacity will be applied regardless.
Or it is just pure control of finances ? To control where the Ri will be counted? Flexible everywhere in group capacity prio only on Vms with the rigt size? Nothing more?
Ok I see. Now is better. There is one thing connected with this and capacity. I saw in QA that the capacity isn't commited. Now it looks that yes. How the capacity is reserved? I wonder if in whole region or only in some zone?
The capacity you select should be reserved for you. It is not a specific node per say but we make sure that there is capacity in that size for you to deploy to somewhere in that region.
As per which zone, I am not aware of this being tied to any specific zone in a region. We will simply ensure you have capacity to deploy in one of the selected zones in a region for the VM size you select.
@yashesvi @Drewm3 feel free to add any other details if you feel I missed any information.
Thanks very much. To close this I have last thing on the list. When I create ds2v3 ri for 1 year and ds2v3 for 3years all flex and order server ds4v3 will be this one covered by both first year?
The reserved instance discount will only apply when you select the DS2V3 series. The only time it would apply to a DS4V3 would be if for some reason there was no DS2V3 series available. So simply picking a DS4V3 series will not mean you get the discount because you have a reserved instance for DS2.
When a series is not available and you have a reserved instance running a specific size you will pick that size in the portal and if it is not available on the backend we will automatically find a different size for you and apply the discount. You won't get to pick the different size yourself. It is handled by the backend.
I understand that I don't specify on which Vm the discount will be applied.
It was ment like this.let say In my subscription I don't have anymore Ds2v3 and I have only one Ds4v3. I assume that system will not find same size therefore it will look for different from the same group. In this case Ds4v3. How the discount wil be applied? For scenario described up?
You won't personally look for any other sizes. You will still select the size you reserved and if it is not available the platform will pick a similar size and apply the discount. At which point you would be charged the discount rate you payed for the original size but applied to the size the platform auto selected for you.
I thought that the flexible allows me change the size. For example when I buy reserved ds4v3. And deploy ds4v3. After some time I stop ds4v3 and start 2x ds2v3 the system will apply the discount on those two based on factor if there isn't ds4v3
@axatix ah yes sorry I misread that portion. Apologies.
@axatix thanks for linking the other issue. Lets close this one out and continue on the other one as I will need to check offline with some product group members to confirm. Will reply shortly to that other issue.
@MicahMcKittrick-MSFT Capacity is not reserved. Reserved instances don't provide a capacity assurance or guarantee and doesnt physically reserve the specified capacity in the data center. If a user who has reserved instance with capacity priority experiences a deployment failure due to capacity issues, it causes a high severity issue at our end and we try to provide additional capacity to mitigate by taking steps such as letting users eat into buffers or defragmenting.