Docker

Environment variables

The accepted evn vars are:

• DB_URL the mongodb database url, the url must contains the database name, for example mongodb://localhost/db
• DISABLE_GRAPHIQL disable graphiql endpoint
• HIDE_ERRORS_TRACEBACK disables the console errors traceback, useful in production to not bloat the logs
• GRAPHIQL_DEFAULT_QUERY the query showed in the graphiql panel, default is ''
• GRAPHIQL_DEFAULT_QUERY_FILE_PATH the query showed in the graphiql panel, loaded inside a file
• GRAPHIQL_DEFAULT_JWT the default jwt used in graphiql headers panel
• MONGOKE_BASE_PATH the path where to serve the graphql, useful for graphiql and reverse proxying
• MONGOKE_CONFIG_URL the url to a configuration file, useful if you can't mount a file inside the container

Compose

The docker image is hosted on docker hub and has name mongoke/mongoke , every time the container is started the python code is generated based on the configuration given at path /conf.yml . When used in docker-compose the code won't be generated unless the configuration changed, to force the regeneration use the command docker-compose up -V When using an external schema configuration you must mount the file inside the container as a volume. An example on a docker-compose file:

version: '3'

services:
    mongoke:
        image: mongoke/mongoke
        volumes:

            - ./confs/testing.yaml:/conf.yml
            - ./confs/schema:/schema

        environment:

            - DB_URL=mongodb://mongo/db

    mongo:
        image: mongo
        logging:
            driver: none

Swarm

When using with docker swarm probably you want to use configs instead of volumes to be able to deploy the containers in multiple nodes:

version: '3'
services:
    mongoke:
        image: mongoke/mongoke
        configs:

            - conf.yml
            - schema.graphql

        environment:

            - DB_URL=mongodb://mongo/db

    mongo:
        image: mongo
        logging:
            driver: none

configs:
    schema.graphql:
        name: domain-${domain_sum}.graphql
        file: schema.graphql
    conf.yml:
        name: mongoke_conf-${mongoke_sum}.yml
        file: mongoke_conf.yaml

Then to deploy the swarm add the env vars for the config versioning (configs in swarm are immutable and their name must change when the config content changes)

mongoke_sum= `cat mongoke_conf.yaml | md5` \
domain_sum= `cat schema.graphql | md5` \
docker --host ssh://your-server.com stack deploy -c stack.yml stackname

Apollo Federation

apollo federation permits you to gllue together different graphql servers together, To add together more graphql servers together you can use the docker image xmorse/apollo-federation-gateway , this container acts as a gateway between your graphql servers and glue together the different schemas. This approach is perfect to add mutations to your existing mongoke service. Keep in mind that when you change one service you have also restart the gateway to make the changes to the federated schema, this can be handled adding a POLL_INTERVAL env variable that makes the gateway scrape the schemas every tot seconds to check if some schema changed. An example using docker-compose:

# docker-compose.yml
version: '3'
services:
    gateway:
        image: xmorse/apollo-federation-gateway
        ports:

            - 8090:80

        environment:
            URL_0: 'http://mongoke/'
            URL_1: 'http://server'
    mongoke:
        image: mongoke/mongoke
        volumes:

            - ./confs/testing.yaml:/conf.yml
            - ./confs/schema:/schema

        environment:

            - DB_URL=mongodb://mongo/db

    server:
        build: server # your custom graphql server to handle mutations and stuff
    mongo:
        image: mongo
        logging:
            driver: none