Skip to content

The CNC configuration file

cnc uses a superset of docker-compose for configuration. Following the conventions set by docker-compose, cnc adds two key x-cnc annotations.

  • One extension is at the root of the document (sibling to services) and is used for app-wide configuration such as resource definitions that don't have a container (e.g. S3 buckets or some message queues).
  • The other is in each service itself. The type key is the most important one. Supported types depend on the included flavor but the two most common are frontend for static site and backend for a container.

You can use an existing docker-compose file and add these annotations, cnc will ignore any services that do not have the x-cnc configuration added.

The goal of this document is to provide:

  • A comprehensive example with comments on each line to help understand the possible configuration.
  • Instructions on how to see the JSON Schema of all the allowed configuration options, which is auto-generated and can be used for tooling and validation as well as understanding.

Example

This is an example of a monorepo (cnc does not actually know or care if in a repo or not, but most code is!) with 2 services, one frontend and one backend, in directories with the same names. cnc is run from the repo root for this example.

x-cnc:
  build_settings:
    platform_settings:
      # see the cloud provider docs for valid types
      machine_type: "E2_HIGHCPU_8"
  resources:
  # resources can be defined here or as services, interchangably
  # some resources e.g. message queue or bucket don't always have a server in dev
  # for valid types and options, see flavor docs
  - name: bucket1
    type: object_storage

services:
  # this will become the service name in CNC
  frontend:
    x-cnc:
      # this is required
      type: frontend
      # within this container, where doe built assets end up?
      # we need to know to copy them to the CDN
      assets_path: dist
      # what command do you run in the container to build the assets
      # this is often different than the docker-compose command to run
      # the dev server
      build: ["yarn", "build"]
      # you can define custom headers for your site
      custom_headers:
      - name: x-frame-options
        value: SAMEORIGIN

    build:
      # for this container, where is the Dockerfile
      # if you don't have one defined, the default is
      # "Dockerfile" in the current working directory
      # if the dockerfile does not exist, included flavors
      # will use nixpacks for the build step to produce a container
      dockerfile: frontend/Dockerfile
      # this can be used to set the docker build context
      context: frontend

  backend:
    command: ["npm start"]

    ports:
    # these are ignored by CNC, your app is expected to listen on $PORT
    # unless you customize the IaC or runtime templates to change behavior
    - "8080:8080"

    deploy:
      # control the resources deployed
      # can also specify different specs for scheduled tasks/workers
      # see worker below for an example
      resources:
        limits:
          memory: 2g
          cpus: 2

      # how many minimum instances of the server to run
      # any autoscaling rules will still apply and can modify this higher
      # see max_scale in system settings below
      # optional, default is 1
      replicas: 1

    build:
      # see comments in frontend above
      context: backend
      dockerfile: "backend/Dockerfile"

    x-cnc:
      type: backend
      # need to define the url path to route to the service
      # unless you customize IaC to subdomain routing
      url_path: /api
      # for DB migrations, what command to run?
      # runs in the container
      migrate: ["prisma", "migrate"]
      # for DB seeding, what command to run?
      # MUST be idempotent!!
      # runs in the container
      seed: ["prisma", "seed"]
      # system holds infra config info
      system:
        # health check is required on AWS apps
        # can be anything that return 200-399 HTTP code
        health_check: /

        # control cloud behavior
        platform_settings:
          # the higher value of min_scale and replicas in deploy will be used as the min
          # optional, default is 1
          min_scale: 1
          max_scale: 4

      workers:
      # will run a daemon without LB connection
      - name: default queue
        # what command to run? will run in same container as the service
        command: ["node", "worker.js", "default"]
        # how many instances of this worker to run
        replicas: 1 # optional, default is 1
        # define resources (will default to service definitions if not defined)
        system:
          cpus: 2
          memory: 4g

      scheduled_tasks:
      # CRON jobs go here
      - name: check task statuses
        command: ["node", "statuscheck.js"]
        # use k8s syntax for all providers/flavors
        # this is every 2 mins (can be expensive)
        schedule: "*/2 * * * *"
        # define resources (will default to service definitions if not defined)
        system:
          memory: 2g

  redis:
    x-cnc:
      # will be merged into resources for the whole app in each environment
      type: cache

    # cnc will ignore this and use cloud service e.g. Elasticache
    image: redis
    restart: always

  db1:
    x-cnc:
      name: db1
      type: database
      # any version supported by cloud e.g. RDS on AWS
      version: 13
      # can override protocol in DATABASE_URL here for cnc-managed vars
      adapter: postgresql

    # cnc will ignore this and use cloud service e.g. Elasticache
    image: postgres:13
    restart: always