omniboard-dev / analyzer

The Omniboard analyzer is a CLI tool that helps you run your Omniboard checks against your source code and then uploads results back to Omniboard for visualization per project, per check results or even using custom dashboards

@omniboard/analyzer

Getting started with Omniboard in less than 5 minutes (video)

Create account, get API key and define checks

1. Create free account for Omniboard.dev
2. Generate API key in the Omniboard.dev (docs)
3. Set API key as an OMNIBOARD_API_KEY environment variable (or pass it in using --api-key flag when running omniboard command, never commit your API key to the version control system)
4. (optional) test your API key using npx omniboard test-connection --api-key <your-api-key> (same as omniboard tc --ak <your-api-key>)
5. Define checks in the Omniboard.dev app

Run in projects

Make sure you have already set OMNIBOARD_API_KEY environment variable in the given environment (CI) or pass it in using --api-key flag when running omniboard command locally

1. install it using npm i -D @omniboard/analyzer in the project we want to analyze (dev dependency)
2. run it using npx omniboard (or run omniboard as a npm script, eg "postbuild": "omniboard"")

or

• npx @omniboard/analyzer (in case it was not pre-installed)

or

• npm i -g @omniboard/analyzer - install it globally to be able to run omniboard in any path without waiting for npx install

Commands

• omniboard analyze - (same as omniboard) analyze project and upload results to Omniboard.dev (and generate local json, optional)
• omniboard batch - Clone (or update) and analyze multiple project repositories and upload results to Omniboard.dev
• omniboard test-connection - test connection to the Omniboard.dev app
• omniboard test-check - test check definition provided as a CLI argument (can be copied from the Omniboard.dev app)

How it works - analyze

1. retrieve current checks defined in the Omniboard.dev app
2. run retrieved checks for the current project (skip checks that are disabled or if project name does not match provided pattern)
3. upload checks results to the Omniboard.dev app (if OMNIBOARD_API_KEY env variable or --api-key flag is present)
4. (Optional) store check results locally (when --json flag was present)
5. Explore results in the Omniboard.dev app using projects, results or dashboards overview

How it works - batch

Setup

1. on the first run, create empty workspace and job file (if not present)
2. (manual step) add repository urls to the job file (in queue array)

Execution

1. batch will clone (or update) repository from the queue
2. batch will run omniboard analyze in the cloned repository
3. batch will remove repository from the queue into done array (or failed array in case of error), you can use --preserve-queue flag to enable multiple runs of the same job file
4. batch will repeat steps 1-3 until every repository from the queue is processed

Global options

Run omniboard --help for list of all supported commands and options (omniboard <command> --help, provides even more details)

• --help - print help
• --verbose - print debug log statements
• --silent - silences the renderer
• --show-check-subtasks - Show checks subtasks in log output (collapsed by default)
• --errors-as-warnings - exit with success (0) even in case of errors and log them as warnings (useful for CI)
• --api-key - pass in API key when not set as an environment variable
• --api-url - pass in URL of the on-prem Omniboard instance (for custom enterprise plans only)
• --json - store data in local json file
• --json-path - location of local json file
• --check-pattern - only run checks matching provided pattern
• --sanitize-repo-url - try to sanitize auth tokens from repo urls

FAQ

Is it possible to run @omniboard/analyzer behind organization proxy?

Yes

The @omniboard/analyzer uses global-agent library which will uses HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables and use them to make requests to https://api.omniboard.dev

Is this uploading my source code to the cloud?

No

The @omniboard/analyzer runs checks against your source code (or even generated artifacts) and uploads results of these checks to the cloud service for further processing. The uploaded content is then just metadata describing the projects and results but NOT the projects themselves.

In theory, a check which matches everything could be constructed but such result will be rejected as the payload would be too large. The limits of how much data can be stored per check and per all checks for a project can be customized in organization settings in the Omniboard.dev app.