How to start traefik dashboard?

[MichelDiz]

In general, I would do:

helm install traefik traefik/traefik --set dashboard.ingressRoute="true"

By digging I tried

helm install traefik traefik/traefik --set="{--dashboard.ingressRoute=true}"

But it won't work at all.

What are the guides here?

[arlyon]

Also struggling with this. Can access traefik (as evidenced by the 404 page) but dashboard is also a 404...

edit: the trailing / is mandatory so ${ip}/dashboard/

[MichelDiz]

The "workaround" I did was clone the repo, edit manually and helm install traefik ./

[ldez]

Hello,

You have to set: dashboard.enable=true and dashboard.ingressRoute=true

https://github.com/containous/traefik-helm-chart/blob/43d63f418836d7aaac03cfca5b54f4287c8acf9c/traefik/values.yaml#L76-L83

FYI this helm chart is now published in this repo: https://github.com/containous/traefik-helm-chart#installing

[MichelDiz]

Hi @ldez , the issue is that I have to clone instead of following the instructions in the readme. To make it work.

[MichelDiz]

You guys could add a new value in https://github.com/containous/traefik-helm-chart/blob/d7fc8d82551dd8c22a842948f32f4dc9233492e7/traefik/templates/dashboard-hook-ingressroute.yaml#L1
with "OR" operator we could do something like

helm install traefik traefik/traefik --set expose.dashboard=true

Just supposing about it, need to test out the hypothesis.

[mmatur]

Hi @MichelDiz,

The dashboard is now deployed by default only on traefik entry-point.

You can do the installation with the following command:

helm install traefik traefik/traefik

The dashboard is accessible with the following command:

kubectl port-forward traefik-xxxx-xxx 9000:9000

https://localhost:9000/

I'm closing this issue because I think the question is answered.

[blasrodri]

@mmatur
I've installed it using

helm traefik treafik/treafik

Then I do:

$ kubectl port-forward traefik-6c4575d8b8-5w2z5 9000:9000

$ curl http://localhost:9000/ping
OK

$ curl http://localhost:9000/
404 page not found

Any idea of what I'm missing?

[mmatur]

@blasrodri

The dashboard is exposed on /dashboard

curl http://localhost:9000/dashboard/

[blasrodri]

Thanks @mmatur . But still having trouble.
Forgot to add:

curl http://localhost:9000/dashboard
404 page not found

[blasrodri]

oh, is the trailing /? haha
Got it, thanks!

[nelsonyaccuzzi]

Hi, I can't get to the dashboard

I've tried with the trailing dash and it doesn't work, but the ping endpoint works

➜ ~ kubectl port-forward $(kubectl get pods --selector "app.kubernetes.io/name=traefik" --output=name -n traefik-v2) 9000:9000 -n traefik-v2
Forwarding from 127.0.0.1:9000 -> 9000
Forwarding from [::1]:9000 -> 9000

➜  ~ curl http://localhost:9000/ping
OK%                                                                                                                              
➜  ~ curl http://localhost:9000/dashboard
404 page not found
➜  ~ curl http://localhost:9000/dashboard/
404 page not found

[yusiwen]

still got 404 when accessing http://localhost:9000/dashboard/
any updates?

[dchernivetsky]

+1 for this. Why is this closed without due investigation.
We have a default chart installation with the following values.yaml:

traefik:
  enabled: true
  additionalArguments: 
    - "--providers.kubernetesingress"
    - "--providers.kubernetescrd.ingressclass=traefik-internal"
    - "--providers.kubernetesingress.ingressclass=traefik-internal"
    - "--metrics.prometheus=true"
    - "--metrics.prometheus.entrypoint=metrics"
    - "--entrypoints.metrics.address=0.0.0.0:9100"
  serviceType: LoadBalancer
  nameOverride: traefik-internal
  fullnameOverride: traefik-internal
  replicas: 1
  rbac:
    enabled: true
  resources:
    limits:
      cpu: 300m
      memory: 200Mi
    requests:
      cpu: 100m
      memory: 200Mi
  autoscaling:
    enabled: true
    minReplicas: 1
    maxReplicas: 20
    metrics:
    - type: Resource
      resource:
        name: cpu
        targetAverageUtilization: 60

It does seem to install the dashboard ingressroute

kubectl get ingressroute -n kube-system
NAME                         AGE
traefik-internal-dashboard   18m

However no dashboard shows up when port-forwarded:

$ curl http://localhost:9000/dashboard/
404 page not found
$ 

EDIT: For anyone else having trouble with this @nelsonyaccuzzi @yusiwen check that your ingressroute for the dashboard has the correct annotation for the ingress class, should look something like this

apiVersion: traefik.containo.us/v1alpha1
kind: IngressRoute
metadata:
  annotations:
    helm.sh/hook: post-install,post-upgrade
    kubernetes.io/ingress.class: traefik-internal
  labels:
    app: traefik-internal
  [...]

Note the kubernetes.io/ingress.class annotation. To be honest this is a poorly designed parameter. If we are expecting users to override the ingress class via a flag to the binary then we can't expect the users to know the effects this flag might have on other resources (ingressRoute in this example).
ingress class should be moved to its own value parameter and modify any and all aspects of the chart that it influences.

TL;DR there's a bug in the chart with the way it handles ingress.class
@ldez @mmatur /reopen

[xrl]

I cleared up this error on my side by tearing down the traefik helm release, reinstalling it. Something is screwed up with the CRDs, as the post directly above me indicates. But I wasn't doing anything to change the ingress classes so those changes didn't help me.

[indigo423]

Ran into a similar issue deploying the Helm chart. What is really important, add the / in the URL when accessing it with port forwarding like here http://localhost:49599/dashboard/.

[NicolasTr]

I ran into a similar issue. I had to add the following to the helm values to be able to port forward to the dashboard:

ingressRoute:
  dashboard:
    enabled: true

[broomfn]

Why do we need the tailing "/" on "/dashboard/", It's really confusing, I thought one of the goals of Traefik was (and I quote)

"With Traefik, you spend time developing and deploying new features to your system, not on configuring and maintaining its working state."

I've spent hours trying to get this working, and things like having to add tailing backslashes really don't help :-(

[lzd-1230]

I ran into a similar issue. I had to add the following to the helm values to be able to port forward to the dashboard:

ingressRoute:
  dashboard:
    enabled: true

That's real methods dude... It's weird that after open ingressRoute but without actual ingressRoute resource in k8s when using kubectl get ingressroute --all-namespace.