What characterizes a span in tracing systems?

A span in tracing systems is characterized by having an operation name and timestamps. This key feature allows spans to represent a single operation within a distributed system, providing essential details about the execution context. The operation name identifies the specific task or service being traced, while timestamps capture when the operation started and how long it took to complete. This information is crucial for understanding performance bottlenecks and for monitoring the flow of requests through various services.

In contrast to other elements mentioned, spans are not solely about log data or focused solely on CPU time; rather, they give a broader view of operation execution, tracing the interactions between services and the time taken across those executions. Moreover, spans do not aggregate multiple operations but rather represent individual operations, allowing for detailed analysis of the performance and dependencies in a microservices architecture.