To make these configuration changes, go to Settings > Metrics > OpenTelemetry metrics.
These settings affect your metric dimensions. Modifying any of them will cause your metrics to change, which may have an impact on existing dashboards, events, and alerts that make use of these dimensions. In this case, they will need to be updated manually.
When Add Meter name and version as metric dimensions is turned on, the Meter name (also referred to as InstrumentationScope or InstrumentationLibrary in OpenTelemetry SDKs)
and version are automatically added as dimensions (otel.scope.name and otel.scope.version) to ingested OTLP metrics.
In the section Allow list: resource and scope attributes you can configure which resource and scope attributes to add as dimensions to ingested OTLP metrics.
When the toggle Add the resource and scope attributes configured below as dimensions is turned on, the attributes defined in the list will be added as dimensions to ingested OTLP metrics if they are present in the OpenTelemetry resource or in the instrumentation scope.
Dynatrace defines a set of default attributes that we consider relevant and beneficial to have on all of your metrics. You can modify the defaults and add extra attributes that you want to have in all of your metrics from the OpenTelemetry resource/instrumentation scope.
The attributes configured on the settings page are only added as dimensions to ingested OTLP metrics when the toggle Add the resource and scope attributes configured below as dimensions (see above) is turned on.
Attribute names are case-sensitive and must be configured with the original name, as exported to Dynatrace by the telemetry source and before any possible transformation took place. For example, an attribute named My:attribute will be renamed to my_attribute upon ingestion
(see name rules) but still needs to be configured as My:attribute in the settings page.
Even though you can modify the default attribute list, Dynatrace does not recommend that you change or remove attributes starting with dt.*. Dynatrace uses these attributes to enrich ingested metrics with Dynatrace-specific dimensions.
Dynatrace allows a maximum of 50 dimensions per metric. For details, see Limits. This configuration, along with the deny list documented below, allows to control the number of dimensions to stay within this limit.
See Attribute ingestion for details on how Dynatrace ingests data of different attributes sharing the same name.
In the section Deny list: all attributes you can configure which attributes to drop from all ingested OTLP metrics.
Upon ingest, the Allow list: resource and scope attributes above is applied first. Then, the Deny list: all attributes is applied. The deny list therefore applies to all attributes from all sources (data points, scope, and resource).
You can use the deny list to remove attributes that you deem not relevant for your use case. For example, you can use the deny list to remove high-cardinality attributes that may be added outside your control (e.g., by instrumentation libraries or resource detectors).
Attribute names are case-sensitive and must be configured with the original name, as exported to Dynatrace by the telemetry source and before any possible transformation took place. For example, an attribute named My:attribute will be renamed to my_attribute upon ingestion
(see name rules) but still needs to be configured as My:attribute in the settings page.
Dynatrace does not recommend adding attributes starting with dt.* to the deny list. Dynatrace uses these attributes to enrich ingested metrics with Dynatrace-specific dimensions.
Enabling Advanced OTLP metric dimensions activates advanced OpenTelemetry metric capabilities for metrics ingested into Grail. With this feature enabled, Dynatrace ingests all resource, scope, and data-point attributes as metric dimensions by default, except for attributes on the Deny list: all attributes. This means the Allow list: resource and scope attributes setting no longer applies to Grail-ingested metrics when this feature is turned on.
Meter name and version are always added as dimensions (otel.scope.name and otel.scope.version) for Grail-ingested metrics, regardless of the Add Meter name and version as metric dimensions setting.
Relaxed ingestion limits apply when this feature is enabled. See API limits and validations for details.
When you turn on Advanced OTLP metric dimensions, Dynatrace processes Grail-ingested metrics to take full advantage of Grail's storage and query capabilities. This includes using primary fields from ingested metric data to enrich metric dimensions and improve queryability, making additional fields available to support cost-allocation workflows, and providing improved support for queries on high-cardinality dimensions.
Keep in mind that enabling this feature changes how dimensions are handled for your Grail-ingested metrics. You may see new dimensions appear on your metrics that were previously excluded by the Allow list: resource and scope attributes setting, or dimension keys which were normalized before now ingested as-is. This can increase the cardinality of some metrics and may affect dashboards, charts, alerts, and saved queries that depend on specific dimension sets. After enabling this feature, review your relevant dashboards and alerts to ensure they still work as expected. You can use the deny list to exclude any unwanted high-cardinality attributes.
After you change the setting, it may take a few minutes for it to take effect across your environment.
By default, Dynatrace ingests the histogram's min|max|sum|count values and drops the buckets. However, you can opt in to also ingest the buckets from explicit histograms.
Go to Settings > Metrics > Histograms and turn on Ingest complete explicit bucket histograms,
The setting applies to all explicit bucket histogram metrics in your environment. After you turn it on or off, it may take a few minutes for the settings to be applied.
Once buckets are ingested, they can be found by appending the suffix _bucket to the histogram's metric name. For example, http.server.request.duration_bucket for a histogram metric named http.server.request.duration.