Read a request as a sequence of work across service boundaries. Begin with the full duration and identify which operations overlap, which wait for another operation, and which sit on the critical path. A wide span is a useful question, not proof of a root cause. Compare requests for the same operation and time window before drawing a conclusion. Changes in traffic, cache state or downstream services can change the picture. Keep the environment and request identifiers with your notes so the observation can be revisited. Use the explorer to select a span and inspect its service context. The displayed values are synthetic illustration data, so no production telemetry or customer payload is required.