Not able to see disk metrics in azure monitoring after dcr + ama setup

Hello Abhaya Sai AMIRAPU,

Greetings! Thanks for raising this question in the Q&A forum

I reviewed your scenario and the most common cause for this exact symptom, disk metrics like disk/used space missing after a DCR plus AMA setup, especially when older VMSS deployments still show it fine, is a mismatch between where the DCR is sending the performance counter data and where autoscale actually reads metrics from. Autoscale rules read from the Azure Monitor Metrics store, not from a Log Analytics workspace. If your DCR only has a Log Analytics destination, the counter will show up fine in the Perf table but will never appear as a metric you can select for autoscaling. This is a very frequent gap when DCRs are created quickly through the portal wizard or copied from an older template, since the azureMonitorMetrics destination is easy to miss.

Please go through the following checks in order.

Step 1: Confirm the DCR has an Azure Monitor Metrics destination, not just Log Analytics

Open your Data Collection Rule as JSON view in the portal, or run:

az monitor data-collection rule show --name <dcr-name> --resource-group <rg-name> --query "destinations"

You should see an azureMonitorMetrics block similar to this, in addition to any logAnalytics block:

"destinations": {
  "azureMonitorMetrics": {
    "name": "myAmMetricsDest"
  }
}

If this block is missing, that alone explains why the metric is not showing where you need it for autoscaling. You will need to add it and reference it in the dataFlows section for the performance counter data source.

Step 2: Confirm the performance counter data source is flowing to that metrics destination

In the same DCR JSON, check the dataFlows array and confirm the stream containing your disk counters lists azureMonitorMetrics as one of its destinations, not only Custom-Perf or a Log Analytics stream.

Step 3: Verify the exact counter definition

For disk used space specifically, confirm the DCR's performanceCounters section includes the LogicalDisk object with the counter you expect, for example:

\LogicalDisk(*)\% Free Space
\LogicalDisk(*)\Free Megabytes

A counter typo or a narrowed instance filter (using a specific drive letter instead of the wildcard) is a common reason newer DCRs quietly collect less than older ones that were set up manually.

Step 4: Check the Metrics blade with the correct namespace

In the portal, go to your VMSS resource, select Monitoring then Metrics, and change the Metric Namespace dropdown to Azure Monitor Metrics (sometimes shown as Custom Metrics or Azure Monitor Agent) rather than the default platform namespace. Guest level counters collected by AMA only appear under this namespace, never under the standard host level Virtual Machine Scale Sets namespace.

Step 5: Confirm the AMA extension version on the instances

Recent AMA extension versions have changed default behavior around custom metric publishing in a few releases. On an affected instance run:

Get-WmiObject Win32_Process -Filter "name = 'MetricsExtension.Native.exe'" | Select Name, CommandLine

Confirm the process is running with -TokenSource MSI and that the managed identity on the VMSS instances is enabled. If this process is not running, metrics will never reach the Azure Monitor Metrics store even if the DCR is configured correctly.

Step 6: Rule out the Azure Diagnostics extension deprecation if this VMSS was ever using WAD

If any of your earlier working deployments were originally built on the classic Azure Diagnostics extension (WAD) rather than AMA, note that WAD was deprecated on March 31, 2026. If your newer VMSS deployments were provisioned from a template that still assumed WAD-style metric names such as disk/used space, that naming convention does not carry over to AMA, since AMA publishes counters using the LogicalDisk performance counter names shown in Step 3, not the WAD metric naming scheme. Double check that your autoscale rule and any dashboards are referencing the AMA counter name, not the old WAD metric name.

Step 7: If everything above is correctly configured and still failing, escalate

If the DCR destinations, dataflows, counter definitions, and extension health all check out and the metric is still missing, open a support case through the Azure portal under Help + Support, since this would indicate a service side ingestion issue rather than a configuration problem, and include the DCR resource ID and the affected VMSS instance IDs for faster triage.

If this answer helps you kindly accept the answer which will help others who have similar questions.

Best Regards,

Jerald Felix.