Getting Started with Nuclio on Kubernetes

On This Page

Follow this step-by-step guide to set up a Nuclio development environment that uses a Kubernetes cluster.


Before starting the set-up procedure, ensure that the following prerequisites are met:

Install Nuclio

At this stage you should have a functioning Kubernetes cluster, a Docker registry, and a working Kubernetes CLI (kubectl), and you can proceed to install the Nuclio services on the cluster (i.e., deploy Nuclio). For more information about kubectl, see the Kubernetes documentation.

Create a Nuclio namespace by running the following command:

All Nuclio resources go into the "nuclio" namespace, and role-based access control (RBAC) is configured accordingly.
kubectl create namespace nuclio

Create a registry secret: because Nuclio functions are images that need to be pushed and pulled to/from the registry, you need to create a secret that stores your registry credentials. Replace the <...> placeholders in the following commands with your username, password, and URL:

If you want to use Docker Hub, the URL is

read -s mypassword
<enter your password>

kubectl create secret docker-registry registry-credentials \
    --namespace nuclio \
    --docker-username <username> \
    --docker-password $mypassword \
    --docker-server <URL> \

unset mypassword

Create the RBAC roles that are required for using Nuclio:

You are encouraged to look at the nuclio-rbac.yaml file that's used in the following command before applying it, so that you don't get into the habit of blindly running things on your cluster (akin to running scripts off the internet as root).

kubectl apply -f

Deploy Nuclio to the cluster: the following command deploys the Nuclio controller and dashboard, among other resources:

kubectl apply -f
In this example, the Nuclio dashboard service has full access to the local machine's Docker daemon. If you're concerned about the security implications, isolate the dashboard in its own node. The Nuclio team is working with the community to establish a secure and robust on-cluster build mechanism.

Use the command kubectl get pods --namespace nuclio to verify both the controller and dashboard are running.

Forward the Nuclio dashboard port: the Nuclio dashboard publishes a service at port 8070. To use the dashboard, you first need to forward this port to your local IP address:

kubectl port-forward -n nuclio $(kubectl get pods -n nuclio -l -o jsonpath='{.items[0]}') 8070:8070

Deploy a function with the Nuclio dashboard

Browse to http://localhost:8070 (after having forwarded this port as part of the Nuclio installation) to see the Nuclio dashboard. Select the "default" project and then select New Function from the action toolbar to display the Create function page (http://localhost:8070/projects/default/create-function). Choose one of the predefined template functions, and select Deploy. The first build populates the local Docker cache with base images and other files, so it might take a while to complete, depending on your network. When the function deployment completes, you can select Test to invoke the function with a body.

Deploy a function with the Nuclio CLI (nuctl)

Run the following Nuclio CLI (nuctl) command from a command-line shell to deploy the example helloworld Go function. If you're using Docker Hub, the URL should include your username —<username> — and you might also need to log into your Docker Hub account (docker login) on the installation machine before running the deployment command. You can add the --verbose flag if you want to peek under the hood.

nuctl deploy helloworld \
    --namespace nuclio \
    --http-trigger-service-type nodePort \
    --path \
    --registry <URL>
The command in the previous code snippet exposes the function externally using a nodePort. This is done for demonstration purposes only. For more information about exposing your function, see Exposing a function.

When the function deployment completes, you can get the function information by running the following CLI command:

nuctl get function helloworld

Sample output -

  nuclio    | helloworld  | default | ready |     42089 | 1/1   

You can see from the sample output that the deployed function helloworld is running and using port 42089.

Because the function is exposed using a nodePort, you can run the following CLI command to invoke it:

nuctl invoke helloworld --method POST --body '{"hello":"world"}' --content-type "application/json"

Sample output -

> Response headers:
Server = nuclio
Date = Thu, 18 Jun 2020 06:56:27 GMT
Content-Type = application/text
Content-Length = 21

> Response body:
Hello, from Nuclio :]

What's next?

See the following resources to make the best of your new Nuclio environment: