Query Regional Core Usage by Resource

NPizzuti 56

I very frequently run in to the "Operation could not be completed as it results in exceeding approved Total Regional Cores quota" error. I understand what this means and how to increase my subscription's limit, but first I would like to investigate what services/resources are using all the vCPUs.

How can I query for the resources or services currently using vCPUs?

Documentation reviewed that was not helpful:
https://zcusa.951200.xyz/en-us/azure/virtual-machines/windows/quotas (Covers how to query for the overall usage and limit, I want a breakdown on what is contributing to usage)
https://zcusa.951200.xyz/en-us/azure/azure-portal/supportability/regional-quota-requests (I do not wish to increase my quota at this time)

Mike Urnun 9,831 Reputation points Microsoft Employee

2022-06-22T20:50:58.003+00:00
Hello @NPizzuti - Thanks for posting on MS Q&A. You can use Azure Log Analytics to troubleshoot the exact Processes that may be overutilizing your VM resources. Please review the following post: https://zcusa.951200.xyz/en-us/answers/questions/113153/cpu-utilization-logs-location.html

Some links that might help:

https://blog.zuehlke.cloud/2020/12/monitor-processes-on-azure-vms-with-log-analytics/

https://www.youtube.com/watch?v=4AW1cdnz3ac
NPizzuti 56 Reputation points

2022-06-23T14:33:53.83+00:00

Thanks for the response @Mike Urnun - unfortunately, I am not looking for high individual CPU usage, I'm looking for high vCPU allocation. Essentially a breakdown of what services are using how many of the "Total Regional vCPUs" detailed on this page: https://zcusa.951200.xyz/en-us/azure/virtual-machines/windows/quotas. I apologize if I misunderstood the resources you posted, it seems that all of them are focused on finding high percentage usage of individual CPUs.
NPizzuti 56 Reputation points

2022-06-23T15:55:22.43+00:00

For example, if my subscription's limit is 16 vCPUs and I have two Azure Databricks clusters in different workspaces, one using 8 vCPUs and one using 4 vCPUs, I would expect to see:
Total Regional vCPUs: 12 16 Count (as provided by Get-AzVMUsage -Location "East US"), but what I would like to receive information about is:
Azure Databricks Workspace 1 (or Cluster 1): 8
Azure Databricks Workspace 1 (or Cluster 2): 4
Mike Urnun 9,831 Reputation points Microsoft Employee

2022-06-23T20:53:21.947+00:00

@NPizzuti Thanks for providing additional information. I'll look into it deeper and share my findings with you soon.
David Bird 1 Reputation point

2022-09-01T08:47:21.943+00:00

This is the exact question I have too!
I would like to know which resources are being used by my regional quota.
I do not want to know how hard they are working - I need to know what they are so I can delete some of them.
Please help :)
David Bird 1 Reputation point

2022-09-01T09:21:17.577+00:00

Get-AzVMUsage -Location "germanywestcentral" | Where-Object {$_.CurrentValue -gt 0}

This is good - but I need to know WHICH resources are using those vCPUs
Stuart Cary 5 Reputation points

2023-08-10T14:42:56.84+00:00

Hello, would like to bump this thread back up. We are running into the same issue. There is no way to understand which resources are contributing to regional vCPU limit. Please advise
Chris Sheridan 0 Reputation points

2024-01-13T19:45:30.11+00:00

I agree as well. I hit the quota limit but I only have 2 VMs and 2 Kubernets clusters.
Alex A. Cordeiro 0 Reputation points

2024-01-18T13:31:57.79+00:00

same here, any update?
Kasper Heyndrickx 0 Reputation points

2024-03-14T07:59:09.7833333+00:00

I have the same issue...
Sebastian Bargas 0 Reputation points

2025-01-06T14:19:51.4866667+00:00

I have the same problem. However, I noticed that if you use Kubernetes, each job node will consume the total VCPUs of the chosen size.

For example:

If your Kubernetes service uses a Standard_B2s node size (2 cpus), and you have 3 nodes, in total 2x3 = 6 VCPUs will be being used.

1 answer

Vidhya Sagar Karthikeyan 396 Reputation points

2025-01-08T13:15:38.8+00:00

This informatin should definitely be available in the azure billing details. Go to Cost Management --> Billing, you should be able to see all the resources that's used/provisioned and billed. This should give you a good idea of whats been used. You can also export the cost data to the lake and do the analysis there.

You can also get an high level view from Cost Analysis sectin in the portal which should be able to provide you details about which service, service type, etc. Based on the info you can further go into the resource groups and figure out why its been used. For instance a databricks instance max worker thread is set to a large number.

If you are using Databricks, you can turn on biling schema . This will create a new schema called system . This will start recording every single cluster and number of master and worker nodes created with date and timestamp. You can then later query the data to know which cluster was using high number of worker nodes. There are few tables which provide you the details, for example select * from system.billing.usage
Please sign in to rate this answer.

0 comments No comments
Sign in to comment

Use comments to ask for clarification, additional information, or improvements to the question.

Share via

Query Regional Core Usage by Resource

1 answer

Your answer