Metrics

GET 
/api/gateway/metrics

KPIs for the OpenAI LLM Monitoring dashboard: volumes by decision/stage/ agent/model, tokens, latency, and an hourly time-series over the window.

Reads the SAME facts and definitions as /api/traffic (shared/traffic_facts.py): an unknown value is NULL, never 0 - tokens are summed over model calls that reported usage only (NULL when none did), latency over measured completed calls only. allowed / blocked / gated / errors are request OUTCOMES (allowed + blocked + gated + errors = total; an allowed request the upstream answered with HTTP >= 400 is an error).

Request

Responses

200

Successful Response

401

The token is missing, invalid or expired. Refresh it, or sign in again.

403

Your role does not allow this. The detail names the missing permission or role.

422

Validation Error