Deploying Streamlit app on Google Cloud Run

Last updated Aug 17, 2023 Edit Source

• Deployment
• Streamlit
• Google cloud run
• Serverless deployment

• A Google Cloud account
• gcloud CLI installed on local system.
• Docker installed on local system.
• Streamlit application to be deployed

Google Cloud Run runs containerised applications. To prepare Streamlit app for deployment, we will need to create a Docker container. Following are the steps:

1. Create a Dockerfile: Crafting a Dockerfile that specifies the environment and dependencies required for the Streamlit app.

1
2
3
4
5
6

FROM python:3.8  
EXPOSE 8080  
WORKDIR /app  
COPY . ./  
RUN pip install -r requirements.txt  
ENTRYPOINT ["streamlit", "run", "main.py", "--server.port=8080", "--server.address=0.0.0.0"]

2. Building the docker image: Using the docker command line tools, we can build the Docker image of Streamlit app.

1

docker build -t <image_name> .

3. Testing the docker container: We can run the docker container locally to ensure that our Streamlit app works as expected within the containerised environment.

1

docker run -p 8080:8080 <image_name>

1. Open Google Cloud Console: Access the Google Cloud Console using the GCP account.
2. Enable the Cloud Run API: If you haven’t already, enable the Cloud Run API for the GCP account.
3. Deploy using Cloud Console: Following the Cloud Run Interface to deploy the dockerised streamlit app. We can also set environment variables and CPU limitations.

1

gcloud builds submit --tag asia.gcr.io/<PROJECT_ID>/<RANDOM_PROJECT_NAME> 

1. Search for Container Registry and then you would be able to see the recent container image deployed.

2. Click Deploy and select Deploy to Cloud Run to deploy the app.

3. The app deployment will start and take some time to finish

After deploying the app, we can test and monitor its usage:

1. Access the deployed app: Once, the deployment is complete, we receive a URL where the Streamlit app is accessible.
2. Monitoring and Troubleshooting: Using the Google Cloud Console, we can monitor the app. This includes checking logs, monitoring usage, and identifying any issues that might arise.

For effecient app updates, we can implement CD:

1. Setting up source control: We can link our Github code repository to Google Cloud Build for automated deployments
2. Configure Triggers: We can configure triggers to build and deploy our app whenever changes are pushed to the repository