hasura

Hasura

Hasura Cloud configuration providing a unified GraphQL API over our datastores for the Catalog application.

Local Development with Docker

Prerequisites

• Docker Desktop installed and running
• Hasura CLI installed
• Doppler CLI installed
• Environment variables configured (see the top-level README)

Running Locally

1. Set the NPM_TOKEN environment variable (retrieve from Doppler):

   export NPM_TOKEN=*npm_token*

2. (Optional) Clear old data if needed:

   yarn dcs-local

3. Start Hasura:

   yarn dcu-local  # Wait ~15s for full initialization (may prompt for Heroku auth)
   yarn hasura-init
   cd hasura/
   hasura console  # Opens Hasura Console in your browser

Making Schema Changes

Always make data model changes locally first, then apply as migrations to production.

1. Edit locally: Changes made in hasura console automatically generate migration files in hasura/migrations/
2. Review: Create a PR with your changes
3. Apply to production after merging:

   hasura migrate apply --endpoint https://catalog-prod.hasura.app --admin-secret '*secret*'
   hasura metadata apply --endpoint https://catalog-prod.hasura.app --admin-secret '*secret*'

Note: The admin secret must be surrounded by single quotes. Find it in Doppler under catalog/prd as CATALOG_PROD_HASURA_ADMIN_SECRET.

Stopping Hasura

Two options for stopping local containers:

• yarn dcs-local — Stops containers and removes Docker images/volumes. Use this to ensure fresh container instances on next startup.
• yarn dcs — Stops containers only, preserving images/volumes/containers.

Note: Database backup data persists in hasura/Docker/docker-entrypoint-initdb.d/latest.dump, so yarn dcu-local typically only needs to run once per branch initialization.