Reference · Glossary

Span (tracing)

Last updated Jul 17, 2026

One **timed unit of work** in a trace — e.g. embed query, retrieve chunks, call chat completion.

#When to use

Debugging multi-step AI apps where you need latency and inputs per step.

Single-call demos with no observability requirement.

Trace shows `retrieve_docs` span returned 0 chunks in 12ms → root cause before blaming the LLM.