Deploy Streamlit using Kubernetes

So you have an amazing app and you want to start sharing it with other people, what do you do? You have a few options. First, where do you want to run your Streamlit app, and how do you want to access it?

  • On your corporate network - Most corporate networks are closed to the outside world. You typically use a VPN to log onto your corporate network and access resources there. You could run your Streamlit app on a server in your corporate network for security reasons, to ensure that only folks internal to your company can access it.
  • On the cloud - If you'd like to access your Streamlit app from outside of a corporate network, or share your app with folks outside of your home network or laptop, you might choose this option. In this case, it'll depend on your hosting provider. We have community-submitted guides from Heroku, AWS, and other providers.

Wherever you decide to deploy your app, you will first need to containerize it. This guide walks you through using Kubernetes to deploy your app. If you prefer Docker see Deploy Streamlit using Docker.

  1. Install Docker Engine
  2. Install the gcloud CLI

If you haven't already done so, install Docker on your server. Docker provides .deb and .rpm packages from many Linux distributions, including:

Verify that Docker Engine is installed correctly by running the hello-world Docker image:

sudo docker run hello-world
star

Tip

Follow Docker's official post-installation steps for Linux to run Docker as a non-root user, so that you don't have to preface the docker command with sudo.

In this guide, we will orchestrate Docker containers with Kubernetes and host docker images on the Google Container Registry (GCR). As GCR is a Google-supported Docker registry, we need to register gcloud as the Docker credential helper.

Follow the official documentation to Install the gcloud CLI and initialize it.

We need to create a docker container which contains all the dependencies and the application code. Below you can see the entrypoint, i.e. the command run when the container starts, and the Dockerfile definition.

Create a run.sh script containing the following:

#!/bin/bash APP_PID= stopRunningProcess() { # Based on https://linuxconfig.org/how-to-propagate-a-signal-to-child-processes-from-a-bash-script if test ! "${APP_PID}" = '' && ps -p ${APP_PID} > /dev/null ; then > /proc/1/fd/1 echo "Stopping ${COMMAND_PATH} which is running with process ID ${APP_PID}" kill -TERM ${APP_PID} > /proc/1/fd/1 echo "Waiting for ${COMMAND_PATH} to process SIGTERM signal" wait ${APP_PID} > /proc/1/fd/1 echo "All processes have stopped running" else > /proc/1/fd/1 echo "${COMMAND_PATH} was not started when the signal was sent or it has already been stopped" fi } trap stopRunningProcess EXIT TERM source ${VIRTUAL_ENV}/bin/activate streamlit run ${HOME}/app/streamlit_app.py & APP_ID=${!} wait ${APP_ID}

Docker builds images by reading the instructions from a Dockerfile. A Dockerfile is a text document that contains all the commands a user could call on the command line to assemble an image. Learn more in the Dockerfile reference. The docker build command builds an image from a Dockerfile. The docker run command first creates a container over the specified image, and then starts it using the specified command.

Here's an example Dockerfile that you can add to the root of your directory.

FROM python:3.8-slim RUN groupadd --gid 1000 appuser \ && useradd --uid 1000 --gid 1000 -ms /bin/bash appuser RUN pip3 install --no-cache-dir --upgrade \ pip \ virtualenv RUN apt-get update && apt-get install -y \ build-essential \ software-properties-common \ git USER appuser WORKDIR /home/appuser RUN git clone https://github.com/streamlit/streamlit-example.git app ENV VIRTUAL_ENV=/home/appuser/venv RUN virtualenv ${VIRTUAL_ENV} RUN . ${VIRTUAL_ENV}/bin/activate && pip install -r app/requirements.txt EXPOSE 8501 COPY run.sh /home/appuser ENTRYPOINT ["./run.sh"]
priority_high

Important

As mentioned in Development flow, for Streamlit version 1.10.0 and higher, Streamlit apps cannot be run from the root directory of Linux distributions. Your main script should live in a directory other than the root directory. If you try to run a Streamlit app from the root directory, Streamlit will throw a FileNotFoundError: [Errno 2] No such file or directory error. For more information, see GitHub issue #5239.

If you are using Streamlit version 1.10.0 or higher, you must set the WORKDIR to a directory other than the root directory. For example, you can set the WORKDIR to /home/appuser as shown in the example Dockerfile above.

Put the above files (run.sh and Dockerfile) in the same folder and build the docker image:

docker build --platform linux/amd64 -t gcr.io/$GCP_PROJECT_ID/k8s-streamlit:test .
priority_high

Important

Replace $GCP_PROJECT_ID in the above command with the name of your Google Cloud project.

The next step is to upload the Docker image to a container registry. In this example, we will use the Google Container Registry (GCR). Start by enabling the Container Registry API. Sign in to Google Cloud and navigate to your project’s Container Registry and click Enable.

We can now build the Docker image from the previous step and push it to our project’s GCR. Be sure to replace $GCP_PROJECT_ID in the docker push command with the name of your project:

gcloud auth configure-docker docker push gcr.io/$GCP_PROJECT_ID/k8s-streamlit:test

For this step you will need a:

  • Running Kubernetes service
  • Custom domain for which you can generate a TLS certificate
  • DNS service where you can configure your custom domain to point to the application IP

As the image was uploaded to the container registry in the previous step, we can run it in Kubernetes using the below configurations.

Make sure your Kubernetes client, kubectl, is installed and running on your machine.

For configuring the Google OAuth Client, please see Google Auth Provider. Configure oauth2-proxy to use the desired OAuth Provider Configuration and update the oath2-proxy config in the config map.

The below configuration contains a ouath2-proxy sidecar container which handles the authentication with Google. You can learn more from the oauth2-proxy repository.

Create a YAML configuration file named k8s-streamlit.yaml:

apiVersion: v1 kind: ConfigMap metadata: name: streamlit-configmap data: oauth2-proxy.cfg: |- http_address = "0.0.0.0:4180" upstreams = ["http://127.0.0.1:8501/"] email_domains = ["*"] client_id = "<GOOGLE_CLIENT_ID>" client_secret = "<GOOGLE_CLIENT_SECRET>" cookie_secret = "<16, 24, or 32 bytes>" redirect_url = <REDIRECT_URL> --- apiVersion: apps/v1 kind: Deployment metadata: name: streamlit-deployment labels: app: streamlit spec: replicas: 1 selector: matchLabels: app: streamlit template: metadata: labels: app: streamlit spec: containers: - name: oauth2-proxy image: quay.io/oauth2-proxy/oauth2-proxy:v7.2.0 args: ["--config", "/etc/oauth2-proxy/oauth2-proxy.cfg"] ports: - containerPort: 4180 livenessProbe: httpGet: path: /ping port: 4180 scheme: HTTP readinessProbe: httpGet: path: /ping port: 4180 scheme: HTTP volumeMounts: - mountPath: "/etc/oauth2-proxy" name: oauth2-config - name: streamlit image: gcr.io/GCP_PROJECT_ID/k8s-streamlit:test imagePullPolicy: Always ports: - containerPort: 8501 livenessProbe: httpGet: path: /_stcore/health port: 8501 scheme: HTTP timeoutSeconds: 1 readinessProbe: httpGet: path: /_stcore/health port: 8501 scheme: HTTP timeoutSeconds: 1 resources: limits: cpu: 1 memory: 2Gi requests: cpu: 100m memory: 745Mi volumes: - name: oauth2-config configMap: name: streamlit-configmap --- apiVersion: v1 kind: Service metadata: name: streamlit-service spec: type: LoadBalancer selector: app: streamlit ports: - name: streamlit-port protocol: TCP port: 80 targetPort: 4180
priority_high

Important

While the above configurations can be copied verbatim, you will have to configure the oauth2-proxy yourself and use the correct GOOGLE_CLIENT_ID, GOOGLE_CLIENT_ID, GCP_PROJECT_ID, and REDIRECT_URL.

Now create the configuration from the file in Kubernetes with the kubectl create command:

kubctl create -f k8s-streamlit.yaml

Since you are using the Google authentication, you will need to set up TLS support. Find out how in TLS Configuration.

Once the deployment and the service are created, we need to wait a couple of minutes for the public IP address to become available. We can check when that is ready by running:

kubectl get service streamlit-service -o jsonpath='{.status.loadBalancer.ingress[0].ip}'

After the public IP is assigned, you will need to configure in your DNS service an A record pointing to the above IP address.

forum

Still have questions?

Our forums are full of helpful information and Streamlit experts.