Prometheus OTLP Ingestion

OTLP ingestion in Prometheus

Prometheus 3.0 introduces support for OTLP (OpenTelemetry Protocol) metrics ingestion, enabling direct push of OpenTelemetry metrics to Prometheus without requiring metrics exposure via HTTP. This simplifies configuration and helps consolidate your monitoring setup. Additionally, since high cardinality has traditionally impacted the performance of scraping large server fleets, moving from a pull (scrape) to a push model reduces the impact of cardinality on performance. However, depending on your Prometheus storage backend, cardinality remains an important metric to monitor and manage over time.

Below is an example configuration for pushing all metrics to Prometheus:

config.yaml

telemetry:
  metrics:
    otlp:
      exporters:
        - endpoint: "localhost:9090"
          exporter: http
          path: /api/v1/otlp/v1/metrics
          temporality: cumulative # Important

Docker Compose Example

docker-compose.yml

  prometheus:
    image: prom/prometheus:v3.0.0-beta.1
    command:
      - --web.enable-otlp-receiver
    networks:
      - primary
    extra_hosts:
      # https://medium.com/@TimvanBaarsen/how-to-connect-to-the-docker-host-from-inside-a-docker-container-112b4c71bc66
      - 'host.docker.internal:host-gateway'
    ports:
      - '9090:9090'