User Guide

Shiny server can run with two language : R and recently Python. The shiny-k8s toolkit is also compatible with these two languages. We provide base image with various version of the language.

Setup project

• Setup the project to run on Kubernetes
  • The project in GitLab
    • Option 1: Create a blank project
    • Option 2: Use an existing GitLab project
  • Re-use skeleton from example
  • Store the sources in customly named directory
• Define deploy token
  • To view or modify the visibility of your project
  • Create a token that can read the registry
  • Provide token to the CI
  • Indicate into you kubernetes settings that the registry is private
• Adapting the Dockerfile to your needs
  • Line 1: The base image, with the right version
  • Line 5: Working Directory
  • Lines 8-11: The system dependencies
  • Lines 14-17: The package dependencies
    • Python: requirements.txt
    • R: my_packages_to_install.csv
  • Line 20: Grant shiny to create bookmark state
  • Line 23: Copy your app into the image
  • Line 26: Grant shiny to write output data

Run locally

• How it works
  • Volumes
  • Port
  • Command
• Run your container
• Administrate your container

Run on kubernetes

• Ask for your project to be able to deploy to Kubernetes cluster(s)
  • Dev ? Prod ? What environement do I need?
  • Do I want a public URL accessible without the VPN?
  • What resources (CPU, RAM, storage) do I need?
    • How memory overuse is handled by Kubernetes
    • How to observe my memory usage locally
    • How to estimate quota needs from your app ressources
    • Dev vs Prod quota
  • Ask it!
• Adapte the CI to your project
• Configuring your application in Kubernetes (values.yaml)
  • Private registry and deploy token
  • Nodowntime
  • Autoscaling
  • Adjusting resources
  • Ingress annotations
  • Storage
  • Tunning for the production instance
• Access to logs and quota

Frequently Asked Questions

• FAQ