davewain / scale Goto Github PK

Scale speeds up your automated tests by running them in parallel using Docker. Here are some of its features:

• It automagically takes care of setting up multiple databases for each container to connect to (whether that database is Postgres, Redis or so on).
• It autodetects your project's Dockerfile (or attempts to create one for you) and uses that as the environment for your app
• It records the time taken for each test file and re-balances your test files on each container based on the execution time of the previous run
• It uses TOML to manage configuration -- easier to read and edit that JSON or YAML

A Scale Config file for a Rails app looks like this

This initializes a Dockerfile (if a Dockerfile is present, it doesn't do anything) and a scale.toml file in the root of your project.

It does clever things such as detecting what language the project is and creating a stock Dockerfile, for Rails/Node apps it will look at your dependencies and decide whether you need a Postgres/Redis/ElasticSearch instance and configure your scale.toml accordingly.

This spins up your environment based on your scale.toml file, creates sandboxed databases (whether that is Redis/Postgres while retaining the same schema) and runs tests in containers in parallel. The test files are distributed across these containers based on the time it takes to run each test file.

Accepts a glob pattern and expands that into all files in that glob

This accepts a newline separated list of files and uses the JUnit results which are stored in the file located at global.test_results_path and picks the bucket of tests that correspond to that container "index". Two implicit arguments to this function are

a) parallelism b) the index of the container within who's scope these tests are running

If a JUnit results file is not present at that location, then it distributes files based on the nunber of lines in the file