Calculated metrics for services

  • How-to guide
  • 5-min read
  • Published Mar 30, 2021

Dynatrace automatically captures important metrics for services with no configuration required. You might need additional business or technical metrics that are specific to your application. These metrics can be calculated and derived based on a wide variety of available data within the captured distributed trace. You can also split these metrics by multiple dimensions, for example, a requests attribute or an HTTP method.

  • Only new data is written to calculated metrics; retrospective data is not included.
  • You can have up to 500 enabled calculated metrics per environment and up to 100 enabled calculated metrics per service.
  • Classic calculated metrics support at most 100 dimension values. This is referred to as the "top X" rule, as you can select fewer depending on your configuration. However you choose the top 100 dimension values, the remaining dimensions are aggregated into a single timeseries and the dimension value is accessible through a special remainder dimension. The remainder filter condition allowing you to filter on this remainder dimension.

Steps

  1. Go to Settings > Server-side service monitoring > Calculated service metrics > Create new metric.

  2. Enter the name for the metric. The name and the calc:service. prefix are added automatically to the metric key.

  3. Choose the metric source from the Metric source list.

    • If the source is a request attribute, select the required unit.
    • optional To exclude the data contribution of muted requests, turn on Ignore muted requests.

  4. optional Select the management zone. The new metric will be restricted to data from this zone.

  5. Provide conditions to define which requests are included in the calculation.

    If you provide several conditions, all conditions must be fulfilled to use the metric.

    1. Select Add condition.

    2. Select the attribute to be checked.

    3. Select the operator of the condition.

    4. If needed, specify the reference value.

      Preview shows the list of services to be included to the custom metric and the estimation of DDU consumption.

      Preview only considers management zone and conditions based on service attributes. These attributes are marked with [Service property] in the attribute list.

  6. optional Add dimension to your new metric.

    1. Turn on Split by dimension.
    2. Choose the placeholders to define the dimension.
    3. Enter a dimension name.
    4. Define the top X value limit.
      1. In the Number of top values field, specify the amount of top X values to be calculated for the metric.
      2. From the Value sorting and Value aggregation lists, select the sorting and aggregation of the top X values.

  7. Review the metric source and dimension names. They will be used in the UI and API. Once a metric is created, you can't change them.

  8. Select Save metric.

When a placeholder is not available, you can create a custom placeholder. All custom placeholders must be used in the dimension value pattern, alternatevely you can delete unused custom placeholders.

Extraction method

You have two methods to extract the value for a placeholder:

  • Delimiter-based. In this case, Dynatrace extracts the value by checking the value of the source against a reference value and specified delimiter kind and position.
  • Regex-based. In this case, Dynatrace uses regular expression extraction. To learn how Dynatrace uses regular expressions, see Regular expressions in Dynatrace.

Extraction method for request attributes

Placeholders that are based on request attributes provide three options for value extraction:

  • First: The first occurrence of the attribute is used to extract the value.
  • Last: The last occurrence of the attribute is used to extract the value.
  • Count: The value equals the number of attribute occurrences.

Steps

To create a custom placeholder

  1. While creating or editing a calculated service metric, select Add custom placeholder.

  2. Enter a name for your placeholder. The name will be used in the Dimension value pattern field.

  3. Select the source for the dimension.

  4. Choose the extraction method.

    If the source is a request attribute,

    1. optional Select Use from downstream services checkbox to use child calls as the value source.
    2. optional Restrict the child calls to a particular management zone or service tag.

  5. Select Add.

  6. required Use the newly created placeholder in the dimension value pattern.

Related tags
Application Observability