cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
piotr_szwarc
Dynatrace Guide
Dynatrace Guide

 

If you have missing or wrong metrics:

  • Verify the charts in Azure Portal Monitor with Dynatrace Data explorer.
  • Make sure that you compare the same periods.

  • Make sure you compare data with the same resolution in Azure and Dynatrace (for example, one data point per minute, 12:43-12:44). We recommend that you compare data with a one-minute resolution whenever possible.

  • Make sure that you compare granular data split across as many dimensions as possible (Azure: filter/split by).
  • Make sure that you compare the same type of aggregated data (average with average, minimum with minimum, maximum with maximum).

  • Check aggregation functions. If the detailed data shows no discrepancies, you can assume that the underlying data from Azure was imported correctly. Any differences observed when viewing less detailed, aggregated data can result from the method used to group the data over time (SUM/AVG/MIN/MAX) or over dimensions (for example, merging data from multiple combinations of dimension values into a single data point).
    • Example 1: Five data points representing a one-minute period can be aggregated into a single data point for a five-minute period.

    • Example 2: Data from various regions (such as "region=east-us, responseStatus=400", "region=west-us, responseStatus=200") can be combined into a single data point representing the total number of responses across all regions.
Version history
Last update:
‎09 Nov 2023 02:17 PM
Updated by:
Comments
radek_jasinski
DynaMight Guru
DynaMight Guru

Thanks @piotr_szwarc  for these tips. Especially useful is the point: "Check aggregation functions".

IrynaKudlaienko
Dynatrace Advocate
Dynatrace Advocate

When comparing individual data points between Azure and Dynatrace within the same timeframe, with a one-minute resolution and consistent dimensions, if the data points align, it indicates that the metric capture process is functioning correctly and the data points are accurate.

However, disparities may arise when comparing aggregated data. This occurs because Dynatrace captures specific data points from Azure every minute rather than aggregating data directly from Azure itself. As a result, variations in aggregated data can occur.

It is worth mentioning that when Microsoft sends a singular value, such as a count representing a metric, it's interpreted as a "Total" on the Dynatrace platform. Therefore, the aggregation method on the Dynatrace side should be set to "Sum," even if the corresponding method on the Azure side is "Count."

For example, 

Azure Cosmos DB Account (GlobalDocumentDB)

metric: ext:cloud.azure.microsoft_documentdb.databaseaccounts.totalrequests

For individual data points:

Screenshot 2024-05-15 at 17.17.55.png

For aggregated data:

Screenshot 2024-05-15 at 17.58.28.png