How to deploy Traefik Ingress Controller on Kubernetes using Helm

In this tutorial, you will learn how to deploy a Traefik ingress controller on your Kubernetes cluster. You will learn more about Traefik, its benefits and how to set it up and configure it on Kubernetes using Helm.

We will be using a Platform9 Managed Kubernetes cluster for this tutorial. However, you can use this tutorial to deploy Traefik on any Kubernetes cluster of your choice.

What is Traefik

Traefik Kubernetes Ingress is a modern, full-featured ingress controller for Kubernetes with powerful capabilities out of the box. Traefik installs as one or more pods that represent the Traefik controller, ingress proxies, and mesh proxies in your Kubernetes cluster.

Source: https://traefik.io/solutions/kubernetes-ingress/

Prerequisites

• A Kubernetes cluster.
• A Kubectl installation with your Kubernetes cluster from the step above configured as the primary cluster.
• The helm 3 package manager client installed on your local machine. Follow Helm CLI to install the helm client on your machine

Step 1 - Install via CLI or PMK App Catalog

CLI

Once your Kubernetes cluster is up and running, you need to add the helm repository for Traefik to the cluster.

xxxxxxxxxx

helm repo add traefik https://helm.traefik.io/traefik

xxxxxxxxxx

"traefik" has been added to your repositories

Once you have added the helm repository, perform a generic repository update for your helm repositories to fetch the latest upstream updates for the newly-added traefik chart repository.

xxxxxxxxxx

helm repo update

xxxxxxxxxx

Hang tight while we grab the latest from your chart repositories...

...Successfully got an update from the "traefik" chart repository

Install the Traefik chart, ensuring you are scoped to the desired namespace.

xxxxxxxxxx

helm install traefik traefik/traefik

xxxxxxxxxx

NAME: traefik

LAST DEPLOYED: Thu Apr  1 12:38:31 2021

NAMESPACE: default

STATUS: deployed

REVISION: 1

TEST SUITE: None

Once Traefik is installed, validate that the resources were deployed correctly, e.g. the Traefik service was created, the Traefik pod is running, etc.

xxxxxxxxxx

kubectl get svc -l app.kubernetes.io/name=traefik

xxxxxxxxxx

NAME      TYPE           CLUSTER-IP     EXTERNAL-IP   PORT(S)                      AGE

traefik   LoadBalancer   10.21.12.102   10.0.0.4      80:30416/TCP,443:30724/TCP   2m5s

xxxxxxxxxx

kubectl get po -l app.kubernetes.io/name=traefik

xxxxxxxxxx

NAME                       READY   STATUS    RESTARTS   AGE

traefik-5f896b6cc7-q56xl   1/1     Running   0          2m7s

PMK App Catalog

The PMK App Catalog can simplify installations and allow you to modify the application using the UI instead of working from the command line. We will walk through an installation using the App Catalog below.

Start out by navigating to the Apps section. Select + Add New Repository.

Then we will name the repository, in our example we name it Traefik. The URL is going to be the Helm Repository that hosts the charts we need. For Traefik we will use https://helm.traefik.io/traefik. Save and move on to the next step.

Now the available applications will show up under the App Catalog section. We are going to deploy Traefik. Select Deploy on the Traefik tile. This will take us to the configuration of our application.

In the Deploy Application page we can name our deployment, select the cluster, namespace, version, and modify the values associated with the deployment. We will name the deployment traefik and select the cluster we want to use, then select the default namespace. The version will be set to the latest version as of this post: 10.9.1.

If you want to access the Dashboard for Traefik using an IP address and port 9000 then we will need to modify the values file. The Dashboard will deploy with "traefik" as the default entrypoint, which means we need to update the expose option from false to true. Once modified we will select Deploy.

xxxxxxxxxx

"traefik": {

   "expose": true,

   "exposedPort": 9000,

   "port": 9000,

   "protocol": "TCP"

Step 2 - Expose Traefik Dashboard

By default, the helm chart and PMK App Catalog deploy an IngressRoute for the Traefik dashboard.

xxxxxxxxxx

apiVersion: traefik.containo.us/v1alpha1

kind: IngressRoute

metadata:

  annotations:

    helm.sh/hook: post-install,post-upgrade

  labels:

    app.kubernetes.io/instance: traefik

    app.kubernetes.io/managed-by: Helm

    app.kubernetes.io/name: traefik

    helm.sh/chart: traefik-9.18.1

    ...

  name: traefik-dashboard

  namespace: traefik

spec:

  entryPoints:

  - traefik

  routes:

  - kind: Rule

    match: PathPrefix(`/dashboard`) || PathPrefix(`/api`)

    services:

    - kind: TraefikService

      name: api@internal

By default the entryPoint for traefik is not exposed, which means we'll either need to expose the endpoint - like we did above - or change the endpoint to web and add a Host to the match field. This assumes we want to access the Dashboard via port 80 using a domain instead of an IP address.

xxxxxxxxxx

kubectl edit ingressroute traefik-dashboard

xxxxxxxxxx

apiVersion: traefik.containo.us/v1alpha1

kind: IngressRoute

metadata:

  annotations:

    helm.sh/hook: post-install,post-upgrade

  creationTimestamp: "2022-01-27T20:21:28Z"

  generation: 1

  labels:

    app.kubernetes.io/instance: traefik

    app.kubernetes.io/managed-by: Helm

    app.kubernetes.io/name: traefik

    helm.sh/chart: traefik-10.9.1

  name: traefik-dashboard

  namespace: default

  resourceVersion: "7217"

  uid: 7458ad97-2929-4cc1-94f1-14034d1dd968

spec:

  entryPoints:

  - web

  routes:

  - kind: Rule

    match: Host(`demo.pf9.io`) && PathPrefix(`/dashboard`) || PathPrefix(`/api`)

    services:

    - kind: TraefikService

      name: api@internal

In our example demo.pf9.io is pointing at the LoadBalancer associated with the ingress controller. Since this is all local we are updating /etc/hosts - however you could also update local DNS to achieve the same results. If our LoadBalancer was a Public IP then we could update DNS for our domain using an A record or CNAME depending on your provider. Realistically you would want to set this up with certificates and TLS, however that is beyond the scope of this getting started guide.

xxxxxxxxxx

$ kubectl get svc

NAME         TYPE           CLUSTER-IP    EXTERNAL-IP     PORT(S)                      AGE

traefik      LoadBalancer   10.21.75.43   192.168.86.10   80:32111/TCP,443:30796/TCP   5m20s

xxxxxxxxxx

$ ping demo.pf9.io

PING demo.pf9.io (192.168.86.10): 56 data bytes

64 bytes from 192.168.86.10: icmp_seq=0 ttl=64 time=12.816 ms

64 bytes from 192.168.86.10: icmp_seq=1 ttl=64 time=12.781 ms

If you are having trouble reaching the dashboard on the exposed Ingress route, you can alternatively port-forward the connection via kubectl .

xxxxxxxxxx

kubectl port-forward $(kubectl get pods --selector "app.kubernetes.io/name=traefik" --output=name) 9000:9000

xxxxxxxxxx

Forwarding from 127.0.0.1:9000 -> 9000

Forwarding from [::1]:9000 -> 9000

Conclusion

You have now successfully installed Traefik ingress controller for your Kubernetes cluster.

Interested in more Content?