Containerize apps script code and run it on cloud platforms such as Cloud Run outside the context limitations of the Apps Script IDE.
Bruce Mcpherson continues to expand the capabilities of his gas-fakes library, a tool that has already proven valuable for running Google Apps Script locally on Node.js. In his latest update, Bruce demonstrates how to take this a step further by containerising Apps Script code to run on Google Cloud Run.
For developers familiar with the constraints of the Apps Script IDE, particularly the execution time limits, moving logic to a serverless container environment offers a powerful alternative. With the release of version 2.0.2, gas-fakes now includes a managed configuration for Workspace Domain-Wide Delegation (DWD) enabling secure keyless authentication.
Essentially, this allows developers to package their Apps Script logic into containers, enabling execution on scalable platforms like Cloud Run, free from the constraints of the standard IDE.
Authentication and Service Accounts
One of the friction points in moving from a bound script to a cloud environment is authentication. Bruce highlights that while Application Default Credentials (ADC) work well for local development, a more secure method is required for Cloud Run. The updated gas-fakes CLI simplifies this by handling the service account configuration automatically.
The library supports two primary authentication types:
Domain-Wide Delegation (DWD): Recommended for production environments and cross-platform scenarios, such as Cloud Run or Kubernetes.
Application Default Credentials (ADC): A fallback method primarily for local development.
Containerisation Workflow
For those looking to deploy their own scripts, Bruce’s guide walks through the essentials required to containerise a project.
It is important to note that while the core logic can be written in JavaScript, you will need to manually set up the infrastructure configuration. The guide provides the specific code required for:
The Dockerfile: To mimic the Apps Script runtime environment.
Cloud Build Configuration: A cloudbuild.yaml file to manage the build steps.
To tie it all together, the article includes a deployment script (referred to as deploy-cloudrun.sh in the text) which automates the pipeline. It handles everything from creating the Artifact Registry repository and submitting the build, to monitoring the deployment.
Code Example
To help developers get started, Bruce has provided a dedicated repository containing a working example.
Note on usage: The repository contains the core Node.js logic (example.js) and a deployment helper script (named exgcp.sh in the repo). However, to deploy this successfully, you will need to combine these files with the Dockerfile and cloudbuild.yaml configurations detailed in the main article. Bruce notes there is more to come.
The example.js file illustrates how standard Apps Script services are imported and used within the Node.js environment. By requiring gas-fakes, you can access services like DriveApp or SpreadsheetApp using the exact same syntax you use in the Apps Script editor, bridging the gap between local Node development and the Google Cloud runtime.
Summary
This development opens up interesting possibilities for hybrid workflows where lightweight tasks remain in Apps Script, while heavier processing is offloaded to Cloud Run without needing to rewrite the core logic in a different language. If you are interested in trying this out, Bruce has provided example deployment scripts and a troubleshooting guide to help you get started.
