docker-circleboom-hasura

Docker compose file for Hasura with Caddy as reverse proxy.

Hasura GraphQL Engine is a blazing-fast GraphQL server that gives you instant, realtime GraphQL APIs over Postgres, with webhook triggers on database events, and remote schemas for business logic. Hasura helps you build GraphQL apps backed by Postgres or incrementally move to GraphQL for existing applications using Postgres. Check their website here https://hasura.io

This docker compose file uses these containers:

• Hasura v1.3.3 as GraphQL Engine https://hub.docker.com/r/hasura/graphql-engine
• Caddy v2.2.1 as reverse proxy server https://hub.docker.com/_/caddy

Environment Variables

To deploy, create an .env file in the same directory with docker-compose.yml file at server with your environment variables in it.

POSTGRES_SERVER=your-postgres-db-server-ip
POSTGRES_SERVER_PORT=5432
POSTGRES_USER=postgres
POSTGRES_PASSWORD=your-secure-db-password
POSTGRES_TARGET_DB=db-name-you-want-use-with-hasura

HASURA_GRAPHQL_ADMIN_SECRET=your-secure-hasura-admin-password

CADDY_SITE_ADDRESS=:80

Caddy Config

Caddy config is named Caddyfile and it is pretty straightforward.

Default configuration is

{$CADDY_SITE_ADDRESS} {
  reverse_proxy hasura:8080
}

which publishes FusionAuth web interface on the server's port 80 which is defined in .env file above.

Replace :80 with your domain name to get automatic https via LetsEncrypt. Add your domain to serve only https (recommended)

Don't forget to set your DNS first as Caddy automatically validates and creates certificates for you as soon as it started.

Allow 443/tcp port to outside access before rebuilding your docker-compose. This might be tricky as docker modifies iptables on linux directly https://docs.docker.com/network/iptables/

If you don't want to open your web interface to the public, you could always create a tunnel for it

# sudo ssh -N -L 443:127.0.0.1:443 root@server-public-ip -i ~/.ssh/id_rsa

and visit https://127.0.0.1 after your docker containers up.

Some Docker commands

// Spin up the containers
# docker-compose up -d

// List active containers
# docker ps

// See the logs
# docker logs your_container_name

// Restart all in the compose file
# docker-compose restart

// Restart an individual container
# docker restart your_container_name