What information does a span in Datadog APM contain?

A span is the fundamental building block of a trace, representing a single unit of work or operation. According to Datadog's data model, every span must contain a name (describing the operation, e.g., http.request), a start timestamp, and a duration. Spans also contain span tags, which are key-value pairs that add arbitrary context and metadata, such as the service name, resource name, environment, HTTP method, or customer ID. This combination of name, timing, and contextual tags makes the span a rich source of information.

A. This is incomplete. It omits the span name and tags, which are essential for identifying and adding context to the operation the span represents.

C. An HTTP method and route are common examples of data stored in span tags for a web request, but they are not fundamental components of every span.

D. Trace metrics like throughput and latency are aggregated data points calculated from a collection of spans over time, not information contained within a single span.

1. Datadog Documentation, "Trace & Span," Section: "What is a Span?". Defines a span as having "a start time and a duration" and containing "a name, a service, a resource, and other metadata in the form of tags."

2. Datadog Documentation, "APM Data Model," Section: "Spans". Details the required and optional fields for a span, listing name, start, and duration as required, and meta (tags) as a key component.

3. OpenTelemetry Specification, "Spans," Version 1.24.0, Section: "Span". While not Datadog-specific, it establishes the industry standard that Datadog builds upon, defining a span with a name, start/end timestamps, and attributes (tags).