---
title: "Prometheus Metrics"
method: GET
path: "/metrics"
---

# Prometheus Metrics

`GET /metrics`

Prometheus exposition endpoint for Porter metric-based autoscaling.

`ontora_livekit_session_demand` = waiting + in-flight interview sessions
in this region (same predicate as the KEDA design in deploy/keda/). The
livekit-agent-worker's autoscaler scales on it: one pod per ~10 sessions.
Worker-type services can't be scraped by Porter, so the web API serves the
gauge; a single indexed COUNT per scrape (~15-30s) is negligible load.

## Response `200`

Successful Response

- unknown

---

[API](https://skmtc.dev/ontora/apis/ontora-api.md) · [All operations](https://skmtc.dev/ontora/apis/ontora-api/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/ontora/ontora-api/revisions/7f57037fd94b/schema)