Traefik — pronounced traffic — is "The Cloud Native Application Proxy": an MIT-licensed, Go-based HTTP reverse proxy and load balancer that watches your orchestrator or service registry and generates its own routes, aimed at teams running microservices on Docker, Swarm, Kubernetes, Consul, Etcd, Rancher v2, or Amazon ECS.

What it is

Traefik is an HTTP reverse proxy and load balancer written in Go and published under the MIT licence. It integrates with the infrastructure components already present in a deployment — Docker, Swarm mode, Kubernetes, Consul, Etcd, Rancher v2, and Amazon ECS — and configures itself automatically and dynamically from those sources. The project's own framing is that pointing Traefik at your orchestrator should be the only configuration step required. It ships as a single binary and as an official Docker image at hub.docker.com/r/_/traefik/, and exposes a REST API and a web UI.

The concrete problem it solves is route maintenance. Traditional reverse proxies require that you configure each route connecting paths and subdomains to each microservice by hand. In an environment where services are added, removed, killed, upgraded, or scaled many times a day, keeping those routes current becomes tedious. Traefik replaces that manual route table: it listens to the service registry or orchestrator API and instantly generates routes so microservices are reachable from outside, without further intervention. Manual route configuration remains supported through the File provider for cases that need it.

Key capabilities

• Continuously updates its configuration with no restarts.
• Supports multiple load balancing algorithms.
• Provides HTTPS to microservices by leveraging Let's Encrypt, including wildcard certificate support.
• Implements circuit breakers and retry.
• Is WebSocket, HTTP/2, and gRPC ready.
• Exposes metrics to REST, Prometheus, Datadog, Statsd, and InfluxDB 2.X, and keeps access logs in JSON and CLF formats.
• Reads backends from Docker and Swarm mode, the Kubernetes CRD provider, ECS, and the File provider.
• Presents a clean web UI for inspecting the generated configuration, and exposes a REST API.

Who uses it and how

• Kubernetes teams use the Kubernetes CRD provider to let Traefik derive ingress routing from cluster resources rather than a hand-maintained config file.
• Docker and Swarm mode users rely on the Docker provider, so services register and deregister routes as containers start and stop.
• Teams on AWS use the ECS provider for the same automatic route generation on Amazon's orchestrator.
• Operators needing observability wire the metrics integrations and JSON or CLF access logs into existing monitoring and log pipelines.
• Teams that need bespoke routing combine providers with the File provider to keep manual routes alongside discovered ones.

The project has a large community: 64,895 stars, 6,201 forks, and 939 open issues at the time of writing, with a support forum hosted at community.traefik.io.

Getting started

The documented entry point is the 5-Minute Quickstart in the official documentation, which requires Docker. Alternatively, deploy the single Go binary or the official Docker image.

How it compares

The facts provided name no comparable or paid products, so Traefik stands alone in this registry. The only contrast its README draws is with traditional reverse proxies generally: those require per-route manual configuration, while Traefik generates routes from the orchestrator or registry automatically and updates them continuously without restarts. Where a manually configured proxy would need a route added for every path or subdomain, Traefik treats that work as the thing it exists to remove.

When to use it — and when not to

Traefik fits teams already on a supported source — Docker, Swarm mode, Kubernetes, ECS, Consul, Etcd, or Rancher v2 — where services churn frequently enough that manual route upkeep is real work. A self-hoster runs the binary or container, handles certificate issuance and renewal through the Let's Encrypt integration, and provisions whatever metrics and log backends they intend to feed. It is a weaker fit for small deployments with a handful of routes that change rarely, or for teams outside the supported provider list who would end up configuring everything through the File provider anyway. Two caveats are visible in the facts: the issue count sits near 939, and the project maintains a migration guide from v2 to v3, which signals breaking changes across major versions that operators must plan for.

./traefik --configFile=traefik.toml

docker run -d -p 8080:8080 -p 80:80 -v $PWD/traefik.toml:/etc/traefik/traefik.toml traefik

git clone https://github.com/traefik/traefik