OpenTelemetry App

The primary role that the OpenTelemetry app plays is in making raw OTel span data more useful for use cases such as trace search and service observability. It includes several datasets that are used by the Trace Explorer and the Service Explorer to power those experiences out-of-the-box.

Core datasets provided by the app

Span dataset

Span dataset field	Underlying OTel span data
service_name and Service	service.name in the span’s attributes
service_namespace	service.namespace in the span’s Resource attributes
span_name and Operation	span.name
environment	resource_attributes.deployment.environment
service_version	service.version in the span’s Resource attributes
response_status	span.status.code
status_code	http.status_code, rpc.status_code, grpc.status_code, or rpc.grpc.status_code in the span’s attributes
status_message	span.status.message
span_type	span.kind
duration	span.start_time_unix_nano and span.end_time_unix_nano
attributes	The span’s attributes
resource_attributes	The span’s Resource attributes

Span Event dataset

Span Event dataset	Underlying OTel span data
event_name	event.name
attributes	The event’s attributes
resource_attributes	The Resource attributes on the span containing the event

Span Link dataset

Span Link dataset field	Underlying OTel span data
Linked Span	the link’s span_id
Linked Trace	the link’s trace_id
attributes	the link’s attributes

Service Metrics dataset

Contains a set of span-derived service metrics, which are used in the out-of-the-box views in Service Explorer and also available for querying in the Metric Explorer, custom Dashboards, and Worksheets (namespace ServiceExplorer/Service Metrics).

• The metrics are created by aggregating raw span data into 5 minute buckets for requests, errors, and latency. All percentiles are provided in the latency metrics using t-digest.

• Spans of type Service entry point or root spans are being aggregated.

• Metric dimensions: Service, Operation (i.e., endpoint name), service_namespace, service_version, status_code, status_message, environment

The OpenTelemetry app also creates several other datasets derived from the Span dataset that power the Service Explorer.

Span data used by the Service Explorer

Service Explorer feature	Underlying data from Span dataset that powers it
Service discovery	Existence of spans of type Service entry point or root spans (i.e., spans with no parent)
Database discovery	Existence of spans of type Remote call that have db.* attributes
Service naming	service_name
Database naming	attributes.db.system and attributes.db.name in most cases
Database host	attributes.server.address, attributes.net.sock.peer.name, or attributes.net.peer.name
Database port	attributes.server.port, attributes.net.sock.peer.port, or attributes.net.peer.port
Service mapping: Service-to-service edge detection	when a span of type Remote call in one service has a child span of type Service entry point in a different service
Service mapping: Service-to-database edge detection	existence of spans of type Remote call that have db.* attributes
Service inspection: RED metrics	spans of type Service entry point or root spans
Service inspection: endpoints	Operation of type Service entry point
Service inspection: deployments	service_version
Service inspection: errors	response_status, status_message
Service inspection: exceptions	From the Span Event dataset: attributes.exception.message, attributes.exception.type, and attributes.exception.stacktrace
Service inspection: k8s pods	k8s_pod_name
Database inspection: db operation	span.name
Database inspection: db statements	attributes.db.statement

Services and edges show up on the service map if they have reported data in the last 7 days.