# Is it possible to run dashboard UI locally via docker run, but using api proxied from remote server?

**URL:** <https://discuss.kubernetes.io/t/is-it-possible-to-run-dashboard-ui-locally-via-docker-run-but-using-api-proxied-from-remote-server/26944>\
**Category:** General Discussions\
**Tags:** development\
**Created:** [February 2, 2024, 11:20am UTC](https://discuss.kubernetes.io/t/is-it-possible-to-run-dashboard-ui-locally-via-docker-run-but-using-api-proxied-from-remote-server/26944 "2024-02-02T11:20:15Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![ThePlenkov](https://sea2.discourse-cdn.com/flex016/user_avatar/discuss.kubernetes.io/theplenkov/32/14200_2.png) [@ThePlenkov](https://discuss.kubernetes.io/u/ThePlenkov)\
**Post date:** [February 2, 2024, 11:20am UTC](https://discuss.kubernetes.io/t/is-it-possible-to-run-dashboard-ui-locally-via-docker-run-but-using-api-proxied-from-remote-server/26944/1 "2024-02-02T11:20:15Z")

</div>

Hi!

Our company dropped support for Kubernetes UI leaving us just CLI options.

I would love to ask therefore, is it possible somehow to run dashboard UI ( since is like a web app ) in a separate container on localhost, sharing with it api server via kubectl proxy?

Thank you!

---

<div class="post-metadata">

**Author:** ![timwolfe94022](https://sea2.discourse-cdn.com/flex016/user_avatar/discuss.kubernetes.io/timwolfe94022/32/14319_2.png) [@timwolfe94022](https://discuss.kubernetes.io/u/timwolfe94022)\
**Post date:** [February 4, 2024, 2:39am UTC](https://discuss.kubernetes.io/t/is-it-possible-to-run-dashboard-ui-locally-via-docker-run-but-using-api-proxied-from-remote-server/26944/2 "2024-02-04T02:39:19Z")

</div>

[Courtesy of Google DeepMind]

Yes, it is indeed possible to run the Kubernetes Dashboard UI in a Docker container on your local machine while proxying the API server from a remote Kubernetes cluster. Here’s a high-level approach to achieving this:

### Step 1: Set Up `kubectl` Proxy

First, ensure that `kubectl` is configured correctly to communicate with your remote Kubernetes cluster. Once that’s established, you can start a proxy to the Kubernetes API server using the following command:

shCopy code

```auto
kubectl proxy --port=8080

```

This command makes the Kubernetes API server accessible locally at `http://localhost:8080`. The proxy server forwards requests to the remote Kubernetes API server, effectively acting as a middleman.

### Step 2: Run Kubernetes Dashboard in Docker

To run the Kubernetes Dashboard in a Docker container and connect it to the proxied API server, you can follow these steps:

1. **Pull the Kubernetes Dashboard Docker Image:** You’ll need to pull the official Kubernetes Dashboard image from Docker Hub or any other container registry hosting the image.

shCopy code

```auto
docker pull kubernetesui/dashboard:v2.0.0 # Use the latest version

```

1. **Run the Dashboard Container:** Start a Docker container from the Dashboard image. You’ll need to configure the container to use the host network to easily access the `kubectl` proxy. This is done using the `--network="host"` option in the `docker run` command.

shCopy code

```auto
docker run --rm -it --network="host" kubernetesui/dashboard:v2.0.0

```

Using the host network mode allows the Dashboard running inside the container to access services running on the host machine, including the `kubectl proxy`.

### Step 3: Access the Dashboard

With the `kubectl proxy` running and the Dashboard UI running in a Docker container, you can access the Dashboard through your web browser at the following URL:

bashCopy code

```auto
http://localhost:8001/api/v1/namespaces/kubernetes-dashboard/services/https:kubernetes-dashboard:/proxy/

```

### Step 4: Authentication

To access the Dashboard, you will need an authentication token. You can use a Service Account token or a kubeconfig file for authentication. If you don’t have a Service Account with the necessary permissions, you might need to create one and assign the appropriate roles.

1. **Create a Service Account (if necessary):**

shCopy code

```auto
kubectl create serviceaccount dashboard-admin -n kube-system

```

1. **Bind the Service Account to a Cluster Role:**

shCopy code

```auto
kubectl create clusterrolebinding dashboard-admin --clusterrole=cluster-admin --serviceaccount=kube-system:dashboard-admin

```

1. **Get the Token:**

shCopy code

```auto
kubectl -n kube-system describe secret $(kubectl -n kube-system get secret | grep dashboard-admin | awk '{print $1}')

```

Copy the token and use it to log in to the Dashboard UI.

### Notes

- Ensure that your local firewall or security settings allow traffic between your local machine and the remote Kubernetes cluster.
- Be mindful of security implications, especially when exposing the Kubernetes API server, even if it’s just locally. Ensure proper authentication and authorization mechanisms are in place.
- Versions and configurations might vary, so adjust commands accordingly.

This setup should help you get the Kubernetes Dashboard running locally while interacting with a remote Kubernetes cluster.

---

<div class="post-metadata">

**Author:** ![ThePlenkov](https://sea2.discourse-cdn.com/flex016/user_avatar/discuss.kubernetes.io/theplenkov/32/14200_2.png) [@ThePlenkov](https://discuss.kubernetes.io/u/ThePlenkov)\
**Post date:** [February 4, 2024, 8:05am UTC](https://discuss.kubernetes.io/t/is-it-possible-to-run-dashboard-ui-locally-via-docker-run-but-using-api-proxied-from-remote-server/26944/3 "2024-02-04T08:05:26Z")

</div>

Unfortunately this seems to be AI generated response and I do jot think this approach will work just as is… I think we need to have more parameters for container at least

---

<div class="post-metadata">

**Author:** ![yadutaf](https://avatars.discourse-cdn.com/v4/letter/y/fbc32d/32.png) [@yadutaf](https://discuss.kubernetes.io/u/yadutaf)\
**Post date:** [March 19, 2024, 1:19pm UTC](https://discuss.kubernetes.io/t/is-it-possible-to-run-dashboard-ui-locally-via-docker-run-but-using-api-proxied-from-remote-server/26944/4 "2024-03-19T13:19:28Z")

</div>

This question is more than 1 month old, and maybe you already found a way to achieve your goal. However, I’ll post an answer here since I was not able to find any other answer anywhere else, and it may interest other people 🙂

It is possible to run the Kubernetes dashboard locally using Docker compose. Under the hood, this is how the dashboard development mode works. Indeed, running `make serve` from a clone of [GitHub - kubernetes/dashboard: General-purpose web UI for Kubernetes clusters](https://github.com/kubernetes/dashboard) spawns a local `kind` cluster and starts 5 local Docker containers using Docker compose.

Here is a Docker compose Gist that is largely based on this development environment but uses the official released images from the Docker Hub instead: [Run Kubernetes dashboard locally · GitHub](https://gist.github.com/yadutaf/dcbf8668a102438e7b1ec3ff520d0cb6).

This Gist requires a valid Kubeconfig file. It will use `~/.kube/config` by default, but this can be customized using `KUBECONFIG` environment variable.

For example, to run the dashboard locally using the latest official images and using the default kubeconfig file, you can use:

```auto
curl -s https://gist.githubusercontent.com/yadutaf/dcbf8668a102438e7b1ec3ff520d0cb6/raw/802e339518f0fc1458fb0302624eb7ae3c8e184e/kube-dashboard.docker.compose.yaml \
   | docker compose -f- up \
     --remove-orphans \
     --no-attach gateway \
     --no-attach scraper

```

If for any reason the dashboard can not connect to the cluster (eg: when using a specific auth plugin that is not available in the official images like `gke-gcloud-auth-plugin`), you can additionally spawn a local Kube proxy and use a kube config to connect through the proxy:

```auto
# Start the proxy, allowing connections from Docker's bridge
kubectl proxy --address 172.17.0.1 --accept-hosts='^localhost$,^127\.0\.0\.1$,^172\.\d+\.\d+\.\d$'

# Create a stub kubeconfig to connect through this proxy
cat > kubeconfig.proxy <<EOF
apiVersion: v1
kind: Config
clusters:
- cluster:
    server: http://172.17.0.1:8001
  name: proxy
contexts:
- context:
    cluster: proxy
    namespace: default
  name: proxy-default
current-context: proxy-default
EOF

# And finally start the dashboard using this config file
curl -s https://gist.githubusercontent.com/yadutaf/dcbf8668a102438e7b1ec3ff520d0cb6/raw/802e339518f0fc1458fb0302624eb7ae3c8e184e/kube-dashboard.docker.compose.yaml \
   | KUBECONFIG=kubeconfig.proxy docker compose -f- up \
     --remove-orphans \
     --no-attach gateway \
     --no-attach scraper

```

Nota: the scraper does not work in practice, probably because it can not connect to the real metric source. Therefore, no metric will be available in the dashboard.