hatchet/CONTRIBUTING.md

Contributing

This guide will help you understand how to contribute effectively to the Hatchet project.

Getting Started

New to Hatchet? Start with our Architecture docs to familiarize yourself with Hatchet's core system design.

Then, before contributing, check out the following sections:

• Development Environment Setup
• Pull Requests
• Testing
• Running Locally
  • Example Workflow

Development Environment Setup

Ensure all prerequisite dependencies are installed:

• Go 1.25+
• Node.js v18+
  • We recommend using nvm for managing node versions to match the version defined in .nvmrc
• pnpm installed globally (npm i -g pnpm)
• Docker
• task
• protoc
• Caddy
• goose
• pre-commit
  • You can install this in a virtual environment with task pre-commit-install

We recommend installing these tools individually using your preferred package manager (e.g., Homebrew).

Pull Requests

Before opening a PR, check if there's a related issue in our backlog.

For non-trivial changes (anything beyond typos or patch version bumps), please create an issue first so we can discuss the proposal and ensure it aligns with the project.

Next, ensure all changes are:

• Unit tested with task test
• Linted with task lint
• Formatted with task fmt
• Integration tested with task test-integration (when applicable)

If your changes require documentation updates, modify the relevant files in frontend/docs/pages/. You can spin up the documentation site locally by running task docs. By default, this will be available at http://localhost:3000.

For configuration changes, see Updating Configuration.

Testing

Hatchet uses Go build tags to categorize tests into different test suites. For example, these build tags mark a test as unit-only:

//go:build !e2e && !load && !rampup && !integration

func TestMyUnitOfCode() { ... }

Most contributors should familiarize themselves with unit testing and integration testing.

Unit tests verify individual functions without external dependencies:

task test

Integration tests verify components working together with real dependencies (normally spun up via docker compose):

task test-integration

Note: manual testing is acceptable for cases where automated testing is impractical, but testing steps should be clearly outlined in your PR description.

Running locally

1. Start the Postgres Database and RabbitMQ services:

task start-db

2. Install Go & Node.js dependencies, run migrations, generate encryption keys, and seed the database:

task setup

3. Start the Hatchet engine, API server, and frontend:

task start-dev # or task start-dev-tmux if you want to use tmux panes

Once started, you should be able to access the Hatchet UI at https://app.dev.hatchet-tools.com.

Example Workflow

1. Generate client credentials:

task init-dev-env | tee ./examples/go/simple/.env

2. Run the simple workflow by loading the environment variables from ./examples/go/simple/.env:

cd ./examples/go/simple
env $(cat .env | xargs) go run main.go

You should see the following logs if the workflow was started against your local instance successfully:

{"level":"debug","service":"client","message":"connecting to 127.0.0.1:7070 without TLS"}
{"level":"info","service":"client","message":"gzip compression enabled for gRPC client"}
{"level":"debug","service":"worker","message":"worker simple-worker is listening for actions: [process-message:process-message]"}
{"level":"debug","service":"client","message":"No compute configs found, skipping cloud registration and running all actions locally."}
{"level":"debug","service":"client","message":"Registered worker with id: c47cc839-8c3b-4b0f-a904-00e37f164b7d"}
{"level":"debug","service":"client","message":"Starting to listen for actions"}
{"level":"debug","service":"client","message":"updating worker c47cc839-8c3b-4b0f-a904-00e37f164b7d heartbeat"}

Questions

If you have any further questions or queries, feel free to raise an issue on GitHub. Else, come join our Discord!