Skip to main content

Receiving logs from Traefik

Traefik is a modern HTTP reverse proxy and load balancer. It supports exporting logs to an OpenTelemetry endpoint.

OpenTelemetry OpenTelemetry is a collection of APIs, SDKs, and tools used to instrument, generate, collect, and export telemetry data (metrics, logs, and traces) to help you analyze your software's performance and behavior.

FlowG can be used to gather logs from any OpenTelemetry compatible exporter.

⚠️ Important Notice

FlowG only supports gathering logs for now, and does not support metrics and traces. There is an on-going discussion on GitHub regarding the best way to support metrics and traces in the future.

Feel free to join the discussion if you have any ideas or suggestions.

In this tutorial, we will setup Traefik to export logs to FlowG using the OpenTelemetry protocol (OTLP) over HTTP.

Setting up Traefik​

Run Traefik with the following configuration:

experimental:
  otlpLogs: true

log:
  otlp:
    http:
      endpoint: http://127.0.0.1:5080/api/v1/pipelines/default/logs/otlp
      headers:
        Authorization: Bearer YOUR-PERSONAL-ACCESS-TOKEN

NB: This assumes FlowG is running on the same host as Traefik.

Here is another example using Docker Compose:

services:
  traefik:
    image: traefik:latest
    environment:
      - TRAEFIK_EXPERIMENTAL_OTLPLOGS=true
      - TRAEFIK_LOG_LEVEL=DEBUG
      - TRAEFIK_LOG_OTLP_HTTP_ENDPOINT=http://flowg:5080/api/v1/pipelines/default/logs/otlp
      - TRAEFIK_LOG_OTLP_HTTP_HEADERS_AUTHORIZATION=Bearer YOUR-PERSONAL-ACCESS-TOKEN
    ports:
      - "80:80"

  flowg:
    image: linksociety/flowg:latest
    ports:
      - "5080:5080"

Then run with:

docker compose up flowg
# Create your personal access token and set it in the environment variable
docker compose up traefik

Setting up FlowG​

When FlowG starts for the first time, a default pipeline is created which simply store the logs in the default stream. No further setup is required to receive the logs from Traefik:

Traefik logs in FlowG

What's next?​

You can then process the logs, refine them, parse the message, and/or forward them to a third-party system such as a remote Syslog, Datadog, or even another HTTP Webhook.

See the other guides for more ideas.