Publishing Methods for RStudio Connect#
You can use different methods to publish content to RStudio Connect as an end-user, from simple push-button publishing to complex CI/CD pipelines, which are described in the sections below.
For a high-level administrator overview of deploying content into production, refer to the section on asset deployment approaches.
Some of the methods depend on external systems such as Git repositories and CI/CD systems. You can choose the publishing method that works best for your organization's deployment promotion strategy.
You can publish Shiny apps, R Markdown reports, and Plumber APIs directly from the RStudio IDE to RStudio Connect. You can also publish Jupyter Notebooks to RStudio Connect using a notebook extension.
User documentation on publishing content to RStudio Connect.
This publishing method is useful for data scientists who want to publish content directly from development into production.
During push-button publishing, the RStudio IDE will automatically create, upload, and activate a bundle with your application code and a list of dependencies, then your application will be published to RStudio Connect.
Deploying from Git#
You can configure individual applications within RStudio Connect to deploy from Git repositories and update at regular intervals.
User documentation on publishing Git-backed content to RStudio Connect.
Admin documentation on details for Git-backed content in RStudio Connect.
This publishing method is designed to allow data scientists to publish directly from Git repositories to Connect, and have that content get updated at regular intervals without the need for external CI/CD systems like Jenkins or Travis CI.
This functionality is available in RStudio Connect 1.7.6 and newer versions.
You can configure individual applications outside of RStudio Connect with CI/CD systems (e.g., Jenkins, Bamboo) to publish to RStudio Connect on an ongoing basis.
Solutions for programmatic deployment with RStudio Connect.
This publishing method is useful for administrators who want to configure CI/CD pipelines so that data scientists can work on their applications within version-control systems (e.g., Git) while integrating the deployment process into their existing automation and approval workflows.
This deployment method involves manully creating a bundle with application code and a list of dependencies, then uploading and activating the bundle using API endpoints on the RStudio Connect server. Any updates (commits) to the application will result in a new version of the application being published to RStudio Connect.
This functionality is available in RStudio Connect 1.7.0 and newer versions.