Implement Operational Patterns

There are workflows and motions essential to the successful operations of a data science environment. Early in a data science team’s journey, the fundamentals may include establishing good practices around version control.

As teams mature, they may develop or or adopt operational workflows including working with dev-test-prod environments, deploying with CI/CD, auditing and monitoring, and other DevOps functions.

The articles below offer guidance in learning and adopting these common operational workflows that are typically present in enterprise environments.

Featured

Auditing and Monitoring Connect with the Connect Server API

Using the Connect Server API, administrators can create scheduled reports and dashboards to track and manage the content on their servers.

Deploying to Connect from CI/CD Pipelines

Content can be published to Connect from a CI/CD pipeline by utilizing the Connect Server API. A CI/CD pipeline can utilize the API directly, use a Posit-created GitHub Action…

Adopting operations in your development environment

Working with Git in Posit Workbench

Using Git with JupyterLab

Using Git with RStudio Pro

Using Git with VS Code

Command Line Interfaces (CLIs) for GitHub and GitLab

Git Learning Resources

happygitwithr
Get started using Git with R

Learn Git
Online edition of the text "Pro Git" by Scott Chacon and Ben Straub

Learn Git Branching
An interactive way to learn git

Operational patterns in deployment

Code Promotion to Production

Dev / Test / Prod workflows

A discussion on code promotion and implementing this workflow with different server environments

Methods for deploying content to Connect

Push-button, Git-backed, or Programmatic to support different workflows

Examples: Deploying to Connect from CI/CD Pipelines

Each page in this section has detailed examples how various CI/CD system can be utilized to deploy content to Connect.

Best Practices in Managing Deployed Content

Publishing Workflows for Connect

Discussing workflows to incorporate Dev/Test/Prod progressions, peer review, or CI/CD.

Diffable Data Science: A Demo

An example workflow for management of development and production deployments using git

Automation in Production

Options for Scheduling Data Science Tasks

Auditing and Monitoring deployed content

Using the Connect Server API

Create scheduled reports and dashboards to track and manage the content on their servers.

Monitoring RStudio Team Using Prometheus and Graphite

Utilize Graphite or Prometheus endpoints for metrics and monitoring

Example: Plumber API Logging

This asset shows how to log information about plumber API requests and responses by using preroute and postroute hooks. This allows information about the request and the response to be included in the log entry. The logger package is used to generate the log files.