Monitor gateway traffic
Fabric AI Gateway records the requests that pass through it in analytics, and in traces when they’re sampled, so you can monitor traffic without instrumenting your own services. Use Analytics to see aggregated metrics such as cost, latency, and error rate across a time window. Use Traces to inspect individual requests down to the plugin and upstream level.
The gateway records LLM spans with OpenTelemetry GenAI attributes, such as the token counts in gen_ai.usage.input_tokens and gen_ai.usage.output_tokens. To read them, see Inspect traces.
Choose between analytics and traces
Analytics and Traces appear in two places. In the main sidebar, they show traffic for the current namespace. Under Admin, they show traffic for your whole organization. If you run your own data plane, Gateway health also appears under Admin.
Next steps
- To act on what you find, add rate limits, routing, or transforms. For the steps, see Control traffic with plugins.
- To control which model handles each request, see Route LLM requests across models and providers.

