{"id":37819,"date":"2019-11-05T09:00:42","date_gmt":"2019-11-05T16:00:42","guid":{"rendered":"http:\/\/devblogs.microsoft.com\/premier-developer\/?p=37819"},"modified":"2019-11-01T09:05:03","modified_gmt":"2019-11-01T16:05:03","slug":"azure-vm-and-disk-throttling","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/premier-developer\/azure-vm-and-disk-throttling\/","title":{"rendered":"Azure VM and Disk Throttling"},"content":{"rendered":"

App Dev Manager Omer Amin<\/a> describes an improved approach for monitoring disk throttling in Azure virtual machines.<\/p>\n

\n

In my previous blog article (Azure Ultra Disk Storage is here<\/a>), I described a solution for monitoring disk throttling. After getting feedback from customers, I found that the performance was quite slow if you have many virtual machines. It took a lot of time to iterate through the performance data to find instances where the disk or virtual machines was throttled.<\/p>\n

To reduce run times, I found that I could drastically reduce end-to-end time by relying on Log Analytics to perform the work. This did require an additional step to upload the Disk and VM Throttling limits to a custom log in Log Analytics. However, the payload was so small that it did not really impact the total time. Log Analytics does require some additional setup time<\/a> when data is first uploaded to a log. This means that running this script the first time may take up to 20-30 minutes to complete. This is a one-time delay and subsequent run times do not have this delay. This limitation was mitigated by adding a loop to wait and check every minute for the data to show up. Once the data is ready, the script can proceed and compute the results.<\/p>\n

See the updated script at https:\/\/github.com\/omeramin\/Azure-Disk-Throttling-PS<\/a><\/p>\n

The new version uses the following logic:<\/p>\n

    \n
  1. Get list of all VMs and Disks in subscription.<\/li>\n
  2. Get Azure Compute capabilities for each SKU using Azure REST API.<\/li>\n
  3. Match Disks to each corresponding VM, add in the VM IOPS, Disk Bytes Limit.<\/li>\n
  4. Post Disk\\VM data to Log Analytics.\n