Traefik Learn — Interactive Guide

Observability Overview

Traefik provides comprehensive observability out of the box:

Metrics — Prometheus, OpenTelemetry, InfluxDB, Datadog, StatsD
Access Logs — Structured (JSON) or common log format
Tracing — OpenTelemetry, Jaeger, Zipkin, Datadog
Health Checking — Built-in ping endpoint

Metrics

Prometheus

yaml

metrics:
  prometheus:
    addEntryPointsLabels: true
    addServicesLabels: true
    entryPoint: metrics       # Dedicated entrypoint for metrics
    buckets:
      - 0.1
      - 0.3
      - 1.2
      - 5.0

entryPoints:
  metrics:
    address: ":8082"

Use a dedicated entrypoint for metrics so they're not exposed publicly. Or protect the metrics endpoint with middleware.

Key Prometheus Metrics

Metric	Type	Description
`traefik_entrypoint_requests_total`	Counter	Total requests per entrypoint
`traefik_entrypoint_request_duration_seconds`	Histogram	Request duration per entrypoint
`traefik_router_requests_total`	Counter	Total requests per router
`traefik_router_request_duration_seconds`	Histogram	Request duration per router
`traefik_service_requests_total`	Counter	Total requests per service
`traefik_service_request_duration_seconds`	Histogram	Request duration per service
`traefik_service_server_up`	Gauge	Backend server health (1=up, 0=down)

OpenTelemetry

yaml

metrics:
  otlp:
    endpoint: "otel-collector:4317"
    protocol: grpc
    headers:
      api-key: "my-key"
    addEntryPointsLabels: true
    addServicesLabels: true

Datadog

yaml

metrics:
  datadog:
    address: "127.0.0.1:8125"
    pushInterval: 10s
    addEntryPointsLabels: true
    addServicesLabels: true

Access Logs

Configuration

yaml

accessLog:
  filePath: /var/log/traefik/access.log
  format: json     # json or common
  bufferingSize: 100
  filters:
    statusCodes:
      - "200-299"
      - "400-499"
    retries: true
    minDuration: 10ms
  fields:
    headers:
      defaultMode: keep
      names:
        User-Agent: keep
        Authorization: drop
        X-Api-Key: drop
      redirections: true

Log Format

JSON format example:

json

{
  "ClientHost": "192.168.1.100",
  "ClientUsername": "-",
  "RequestAddr": "example.com",
  "RequestHost": "example.com",
  "RequestMethod": "GET",
  "RequestPath": "/api/users",
  "RequestProtocol": "HTTP/2.0",
  "Duration": 45000000,
  "OriginDuration": 42000000,
  "RouterName": "api-router",
  "ServiceName": "api-service",
  "ServiceURL": "http://10.0.0.1:8080",
  "StatusCode": 200,
  "RequestCount": 42,
  "TLSVersion": "1.3",
  "DownstreamStatus": 200,
  "DownstreamContentSize": 1234,
  "RequestContentSize": 0,
  "RequestLine": "GET /api/users HTTP/2.0",
  "FrontendName": "api-router",
  "BackendURL": "http://10.0.0.1:8080",
  "BackendName": "api-service"
}

Distributed Tracing

OpenTelemetry Tracing

yaml

tracing:
  otlp:
    endpoint: "otel-collector:4317"
    protocol: grpc
    headers:
      api-key: "my-key"
    samplingRate: 0.1       # Sample 10% of requests
    attributes:
      - key: environment
        value: production

Jaeger

yaml

tracing:
  jaeger:
    samplingServerURL: "http://jaeger:5778/sampling"
    samplingType: const
    samplingParam: 1
    localAgentHostPort: "jaeger:6831"
    propagation: "jaeger"
    traceContextHeaderName: "uber-trace-id"

Zipkin

yaml

tracing:
  zipkin:
    httpEndpoint: "http://zipkin:9411/api/v2/spans"
    sameSpan: false
    id128Bit: true
    sampleRate: 1.0

Datadog

yaml

tracing:
  datadog:
    localAgentHostPort: "dd-agent:8126"
    globalTag: "env:production"
    prioritySampling: true

Health Checking

Traefik has a built-in health check endpoint:

yaml

entryPoints:
  ping:
    address: ":8081"

ping:
  entryPoint: ping

bash

curl http://localhost:8081/ping
# OK

The ping endpoint is useful for load balancer health checks and container orchestration probes (liveness/readiness in Kubernetes).

Kubernetes Probes

For Traefik itself in Kubernetes:

yaml

livenessProbe:
  httpGet:
    path: /ping
    port: 8081
  initialDelaySeconds: 10
  periodSeconds: 10

readinessProbe:
  httpGet:
    path: /ping
    port: 8081
  initialDelaySeconds: 10
  periodSeconds: 10

Logging

General Logging

yaml

log:
  level: INFO           # DEBUG, PANIC, FATAL, ERROR, WARN, INFO
  filePath: /var/log/traefik/traefik.log
  format: json          # json or common

Dashboard

Traefik's web dashboard provides real-time visibility:

yaml

api:
  dashboard: true
  debug: true

entryPoints:
  dashboard:
    address: ":8080"

Access at http://localhost:8080/dashboard/ (note the trailing slash).

The dashboard shows your full configuration including routing rules and service endpoints. Always protect it with authentication and an IP allowlist.

Grafana Dashboard

A sample Prometheus query for a Grafana dashboard:

promql

# Request rate by router (requests/sec)
sum by (router) (rate(traefik_router_requests_total[5m]))

# P99 latency by service
histogram_quantile(0.99, sum by (le, service) (rate(traefik_service_request_duration_seconds_bucket[5m])))

# Error rate
sum(rate(traefik_router_requests_total{code=~"5.."}[5m])) / sum(rate(traefik_router_requests_total[5m]))

# Backend server health
sum by (server) (traefik_service_server_up)

Next Chapter

Explore the API & Dashboard and how to manage Traefik programmatically.