Deployment
Deploy your CamML app from within RStudio to shinyapps.io or as a Docker image to Google Cloud Run.
Shinyapps.io#
RStudio has some great docs about shinyapps.io, so check out their documentation to read about how to deploy an app.
Google Cloud Run#
Google also has some good docs about deploying a container, but we'll walk through this one since they don't have an R-specific example.
Step 1#
Install the Google Cloud SDK which includes some command line tools that we'll need.
Step 2#
Navigate to your IDE (I'm using VS Code). Open your project folder (the code template GitHub repo folder) in your IDE (Open Folder in VS Code). Create a PowerShell terminal (Terminal < New Terminal) and type:
where {project name} is the name of the project you created on Google Cloud Console, and {image name} is the name of the image (whatever you want!). This code tells Docker to build an image (docker build) with the Dockerfile in the open folder (that's what the . at the end means) and name it as the latest image called gcr.io/{project name}/{image name}.
Running your Docker image locally
You can run your Docker image locally with Docker Desktop, just make sure to specify the local port as 3838 to match the container port. For local testing, you can name your image something else like:
Using Docker Hub or public container repos
Be careful if you want to upload your image to a public container repository such as Docker Hub - Code for your app contains secret keys that allow access to your Google Account. These keys are ignored by GitHub thanks to the .gitignore but could be accessed if someone downloads your container image.
Step 3#
Push the image to the Google Cloud Container Registry using the following code.
Step 4#
Navigate to Google Cloud Console and select your project. On the left-hand navigation menu, scroll down to Google Cloud Run and select the tab. On the Google Cloud Run tab, select Create Service.
Follow the prompt to name the service and select the region. To Configure the service's first revision, use the default selection of Deploy one revision from an existing container image and click Select to select a container image URL. On the right-hand side of the screen, you will see a sidebar appear of available container options - select the image that you pushed to the registry.
Under Advanced settings, set the container port to 3838, set the memory capacity to 2 GiB, and set the minimum and maximum number of instances to 1.
Click Deploy!
Step 5#
Party! 🎉 Your app is now live!
Resources#
Google Tutorial - Build and deploy a service in other languages