Demo OTel Col Prometheus Exporter

This guide shows how to push metrics from an OpenTelemetry Collector to Prometheus using Prometheus’ remote write receiver. The stack uses Docker Compose to run the Collector, Prometheus, and Jaeger (for traces). You’ll learn how to:

Add Prometheus to Docker Compose and enable the remote write receiver
Provide a Prometheus configuration suitable for receiving pushed metrics
Configure the OpenTelemetry Collector to export metrics via prometheusremotewrite
Start the stack and validate metrics in the Prometheus UI

This demonstration uses Prometheus’ remote write API so the OpenTelemetry Collector can push metrics directly into Prometheus. Use this when you want the Collector to push metrics rather than having Prometheus scrape targets.

Overview

Add a Prometheus service to your Docker Compose file and enable the remote-write HTTP receiver.
Mount a simple prometheus.yml that sets global intervals (no scrape_configs needed if the Collector pushes metrics).
Configure the Collector with a prometheusremotewrite exporter that targets Prometheus at http://prometheus:9090/api/v1/write.
Start the stack and verify metrics at http://localhost:9090.

1) Add Prometheus to Docker Compose

Update docker-compose.yml to include Jaeger and Prometheus services, and configure Prometheus to enable the remote write receiver:

version: "3.8"

services:
  jaeger:
    image: jaegertracing/all-in-one:latest
    container_name: jaeger
    ports:
      - "6831:6831/udp"  # Jaeger agent - thrift compact
      - "6832:6832/udp"  # Jaeger agent - thrift binary
      - "5778:5778"      # Configs
      - "16686:16686"    # Jaeger UI (http://localhost:16686)
      - "14268:14268"    # Collector (HTTP)
      - "14250:14250"    # gRPC

  prometheus:
    image: prom/prometheus:latest
    container_name: prometheus
    ports:
      - "9090:9090"
    volumes:
      - ./prometheus.yml:/etc/prometheus/prometheus.yml:ro
    command:
      - "--config.file=/etc/prometheus/prometheus.yml"
      - "--web.enable-remote-write-receiver"

Ports exposed by the stack:

Service	Purpose / UI	Port (host:container)
Jaeger	UI / tracing	`16686:16686`
Prometheus	UI / metrics	`9090:9090`
Jaeger	gRPC	`14250:14250`
Jaeger	HTTP	`14268:14268`
Jaeger	Agent (UDP)	`6831:6831/udp`, `6832:6832/udp`

Notes:

The --web.enable-remote-write-receiver flag enables Prometheus to accept pushed metrics at /api/v1/write.
We mount a local prometheus.yml to control Prometheus behavior from your project directory.

2) Create Prometheus configuration (prometheus.yml)

If the Collector is pushing metrics to Prometheus, you do not need scrape_configs. Create prometheus.yml beside your docker-compose.yml:

global:
  scrape_interval: 15s
  evaluation_interval: 15s

# No scrape_configs are required if the collector pushes metrics directly.
# The remote-write receiver is enabled via the Prometheus process flag:
# --web.enable-remote-write-receiver

Why no remote_write section? The remote_write stanza in Prometheus is for sending metrics out to remote storage. In this setup, Prometheus acts as the receiver — the HTTP endpoint /api/v1/write — which is enabled by the CLI flag above, not by prometheus.yml. Helpful reference:

Prometheus remote write receiver docs: https://prometheus.io/docs/prometheus/latest/storage/#remote-endpoints-and-storage

3) Configure the OpenTelemetry Collector to export metrics via Prometheus remote_write

Add a prometheusremotewrite exporter to your Collector configuration (example filename: collector-config.yaml or otel-collector-config.yaml) and include it in the metrics pipeline. Example Collector config:

receivers:
  prometheus:
    # default Prometheus receiver config that scrapes instrumented targets
  otlp:
    protocols:
      grpc:
      http:

exporters:
  prometheusremotewrite:
    # Prometheus service hostname as visible from Docker Compose network
    endpoint: "http://prometheus:9090/api/v1/write"

  otlp/jaeger:
    # Forward traces to Jaeger (gRPC)
    endpoint: "jaeger:14250"
    tls:
      insecure: true

  debug:
    # Debug exporter that prints pipeline output to the Collector logs
    verbosity: detailed

processors:
  attributes:
    actions:
      - action: insert
        key: environment
        value: "dev"
  filter:
    # configure metric/log filters as needed
  batch:
    timeout: 10s

service:
  pipelines:
    traces:
      receivers: [otlp]
      processors: [attributes, batch]
      exporters: [debug, otlp/jaeger]
    metrics:
      receivers: [prometheus, otlp]
      processors: [filter, batch]
      exporters: [prometheusremotewrite, debug]
    logs:
      receivers: [otlp]
      processors: [filter, batch]
      exporters: [debug]

Key points:

Exporter name: prometheusremotewrite (this is the Collector exporter that POSTs to Prometheus’ /api/v1/write).
Endpoint: http://prometheus:9090/api/v1/write. When running under Docker Compose, the Collector can resolve the prometheus hostname (service name) on the default network.
Keep a debug exporter if you want to inspect metrics in the Collector logs while simultaneously pushing them to Prometheus.
The otlp/jaeger exporter forwards traces to Jaeger on port 14250.

4) Bring up the stack

Start the services in detached mode:

docker-compose up -d

Verify containers are running:

docker-compose ps

Useful container logs and checks:

Action	Command
Prometheus logs	`docker-compose logs prometheus`
Collector logs	`docker-compose logs <collector-service-name>`
Show running containers	`docker-compose ps`

Prometheus UI: http://localhost:9090
Jaeger UI: http://localhost:16686

5) Validate metrics in Prometheus UI

Open http://localhost:9090.
In the “Expression” input, search for metrics emitted by your application. Example metric names from a sample app:
- current_temperature_fahrenheit_degF
- current_humidity_percentage
Run queries to confirm metrics are visible and being updated.

If metrics are missing, check:

The Collector is running and its metrics pipeline includes the prometheusremotewrite exporter.
The endpoint in the Collector exporter is http://prometheus:9090/api/v1/write.
Prometheus was started with --web.enable-remote-write-receiver and the mounted prometheus.yml is valid.
Services are on the same Docker Compose network so hostnames resolve between containers.

Troubleshooting tips

Use the Collector debug exporter (configured above) to see metric payloads in logs.
Inspect Prometheus logs for errors related to /api/v1/write.
Confirm the Collector and Prometheus containers can reach each other by running a quick curl from within the Collector container:
- Enter the container: docker exec -it <collector-container> /bin/sh
- Test connectivity: curl -v http://prometheus:9090/api/v1/write
Ensure no firewall or host port conflicts exist on 9090.

References and further reading

OpenTelemetry Collector: https://opentelemetry.io/docs/collector/
Prometheus remote write docs: https://prometheus.io/docs/prometheus/latest/storage/
Jaeger: https://www.jaegertracing.io/

This completes the end-to-end setup for pushing metrics from an OpenTelemetry Collector to Prometheus using Prometheus’ remote write receiver.

Watch Video

Practice Lab

Demo OTel Col Filter

OTel Col Connectors

​Overview

​1) Add Prometheus to Docker Compose

​2) Create Prometheus configuration (prometheus.yml)

​3) Configure the OpenTelemetry Collector to export metrics via Prometheus remote_write

​4) Bring up the stack

​5) Validate metrics in Prometheus UI

Watch Video

Practice Lab

Overview

1) Add Prometheus to Docker Compose

2) Create Prometheus configuration (prometheus.yml)

3) Configure the OpenTelemetry Collector to export metrics via Prometheus remote_write

4) Bring up the stack

5) Validate metrics in Prometheus UI