Discrepancy in Azure Virtual Machine Metrics via Azure Monitor REST API

Hello zylker,

It seems that the discrepancies you're experiencing with Azure Virtual Machine metrics via the Azure Monitor REST API may be related to the inherent latency in processing metric data.

Azure Monitor metrics can have ingestion delays of 1–2 minutes or more, depending on the metric and the VM’s region. And when polling too frequently, the data might not yet be available or could be partially aggregated.

Example: You poll at 11:01 AM for data up to 11:00 AM, but Azure Monitor hasn't finished ingesting or aggregating metrics for 11:00 AM, so you get partial or zero values.

And sometimes VMs don’t emit telemetry for a short window due to high CPU load, reboots, or agent issues (e.g., Azure Monitor Agent or Log Analytics agent not fully initialized).

• Additionally, if you're encountering unexpected metric namespaces, it could be due to how the API aggregates and categorizes data. It's important to ensure that your queries are correctly formulated to retrieve the intended metrics.

And if your client-side system (where you're observing different metrics) is not time-synced with Azure's time, it might show mismatched results compared to what's available via the API. so ensure your systems use NTP or synchronize with UTC time correctly.

If you believe the provide information has not helpful, please provide me the below details.

1. An example of the missing/inaccurate metrics that is within the last 5 days
2. The Azure resourceId for which you are attempting to query metrics.
3. The details of the metrics query itself. All of this information can be gathered from the request string of the call itself, or from a screenshot of Metrics Explorer, etc:
   • What timeframe was viewed (such as the time range set in the Azure Portal chart, or the start & end time values supplied to a direct API call)
   • What metric names were queried
   • Which aggregation types were used (Minimum, Maximum, Count, Sum/Total, Average)
   • Which, if any, dimension names and dimension values were used  
4. The output retrieved by the metrics query.
5. The expected output that you believe should have been retrieved by the query.
   • It is important when troubleshooting Metrics queries that we clarify the difference between what the query retrieves and what you expected it to retrieve.
   • This difference can be crucial to identifying if there is unexpected behavior that must involve deeper troubleshooting, or if the issue can be explained by a misunderstanding, which offers us a chance to help you.

Please let me know if you have further queries.

Thanks

Hello zylker,

Just want to check if the above Comment worked for you or else please let us know if any help, we are always here to help whenever you need us.

If the comment is helpful, please click "Upvote it"

Thankyou