You can enrich traces with custom spans to document important steps that are unique to your application. Use quotient.tracing.start_span as a context manager or access the active span via OpenTelemetry utilities.

​
Create a Manual Span

custom_span.py
from opentelemetry.trace import get_current_span

from quotientai import QuotientAI
from quotientai.tracing import start_span

quotient = QuotientAI(api_key="your-quotient-api-key")

quotient.tracer.init(
    app_name="my-app",
    environment="dev",
)

@quotient.trace('support-agent')
def main() -> None:
    with start_span('reranker'):
        span = get_current_span()
        span.set_attribute('reranker.latency_ms', 105)
        span.set_attribute('reranker.model', 'cross-encoder-large')

        # run your custom logic here

if __name__ == "__main__":
    main()

​
Best Practices

  • Give spans descriptive names that identify the step or component (e.g., reranker, sql-query).
  • Set attributes for metrics, inputs, or decisions you care about—anything that helps debug later.
  • Keep spans narrow in scope so they clearly mark a single unit of work.
Next: Vector Database Instrumentation