You are not reading the latest stable version of this documentation. If you want up-to-date information, please have a look at master.

5 - Create Manifest

The Manifest file specifies the details EmbassyOS needs to operate a service. It is the connection point between your service and EmbassyOS.

In this file, values and actions exist for:

  • Displaying the service in the marketplace

  • Specifying the project assets (eg. icon, instructions, license)

  • Defining the docker mount points

  • Specifying how to configure the service

  • Relaying how to run health checks, backups, and other custom actions

  • Outlining dependency relationships (if applicable) and configuration rules for dependencies

  • Denoting copy to display in EmbassyUI elements, alerts, descriptions

This file can be written in:

  • yaml

  • toml

  • json

Let’s create a yaml manifest file for our hello-world project:

touch manifest.yaml

And populate it with the following example manifest (see the line comments for a description of each key and view the full type specification here):

# The package identifier used by the OS
id: hello-world
# A human readable service title
title: "Hello World"
# Service version - accepts up to four digits, where the last confirms to revisions necessary for EmbassyOS - see documentation:
version: 0.3.0
# Release notes for the update - can be a string, paragraph or URL
release-notes: "Upgrade to EmbassyOS v0.3.0"
# The type of license for the project. Include the LICENSE in the root of the project directory. A license is required for a Start9 package.
license: mit
# The Start9 wrapper repository URL for the package. This repo contains the manifest file (this), any scripts necessary for configuration, backups, actions, or health checks (more below). This key must exist. But could be embedded into the source repository.
wrapper-repo: ""
# The original project repository URL. There is no upstream repo in this example
upstream-repo: ""
# URL to the support site / channel for the project. This key can be omitted if none exists, or it can link to the original project repository issues.
support-site: ""
# URL to the marketing site for the project. This key can be omitted if none exists, or it can link to the original project repository.
marketing-site: ""
# The series of commands to build the project into an s9pk for arm64/v8. In this case we are using a Makefile with the simple build command "make".
build: ["make"]
# Minimum required version of EmbassyOS
min-os-version: "0.3.0"
# Human readable descriptors for the service. These are used throughout the EmbassyOS user interface, primarily in the marketplace.
# This is the first description visible to the user in the marketplace.
short: Example service
# This description will display with additional details in the service's individual marketplace page
long: |
    Hello World is a simple example of a service wrapper that launches a web interface to say hello and nothing more.
# These assets are static files necessary for packaging the service for Start9 (into an s9pk). Each value is a path to the specified asset. If an asset is missing from this list, or otherwise denoted, it will be defaulted to the values denoted below.
    # Default =
    license: LICENSE
    # Default = icon.png
    icon: icon.png
    # Default =
    # Default = image.tar
    docker-images: image.tar
# The main action for initializing the service. Currently, the only type of action available is docker.
    # Docker is currently the only action implementation
    type: docker
    # Identifier for the main image volume, which will be used when other actions need to mount to this volume.
    image: main
    # The executable binary for starting the initialization action. For docker actions, this is typically a "" file. See the Dockerfile and the in this project for additional details.
    entrypoint: ""
    # Any arguments that should be passed into the entrypoint executable
    args: []
    # Specifies where to mount the data volume(s), if there are any. Mounts for pointer dependency volumes are also denoted here. These are necessary if data needs to be read from / written to these volumes.
        # Specifies where on the service's file system its persistence directory should be mounted prior to service startup
        main: /root
# Health checks
    name: Web Interface
    description: Ensures the network interface is accessible via HTTP.
    type: docker
    image: main
    entrypoint: "sh"
    args: ["-c", "curl --silent --show-error --fail http://loacalhost:80"]
    # When `inject` is true, the health check will use the main image to run the health check. This is faster as there is no need to spin up an additional docker container
    # When `inject` is false, the health check will use whatever image is specified. This is useful when using a system image with additional utilities to run a health check. If inject=false, then system must equal true
    inject: true
    # Optional if false - indicates if an image that is preloaded onto the system will be used
    system: false
    # Required - valid values are yaml, toml, json
    io-format: json
# Specifies how to get and set configuration file values for the service. NOTE: This stanza can be left empty (null) if the service has no configuration options.
# The config action to run to get the specified config file (default is config.yaml)
        # The type of implementation to run this action (currently, only Docker is available)
        type: docker
        # The Docker image to run the action command in. This could be the service's main image, or an image that is preloaded onto the system, like compat (which holds compatible helper functions for default functionality)
        image: compat
        # Indicates if an image that is preloaded onto the system will be used
        system: true
        # The initial run command to execute the config get action
        entrypoint: compat
        # Any arguments that need to be passed into the run command
        - config
        - get
        - /root
        - "/mnt/assets/config_spec.yaml"
        # The locations at which to mount the specified Docker images
            compat: /mnt/assets
            main: /root
        # Required - valid values are yaml, toml, json
        io-format: yaml
# The config action to run to set the specified config file (default is config.yaml). Details for the keys below are the same as above.
    type: docker
    image: compat
    system: true
    entrypoint: compat
    - config
    - set
    - hello-world
    - /root
    - "/mnt/assets/config_rules.yaml"
        compat: /mnt/assets
        main: /root
    io-format: yaml
# This is a key value map specifying dependent services that this service needs in order to function. The keys are the package id's on which you depend. NOTE: if developing a standalone service, you may leave this stanza as an empty object (the key dependencies is required)
    # Key must be the package id of another service that exists in the marketplace
        # The version range that is acceptable for this dependency
        version: "^"
        # Describes if the dependency is critical to the service functioning. If the dependency is critical, the service will stop if this dependency is stopped.
        critical: false
        # Specifies the requirement type of the dependency
            # "Opt-out" means the dependency will be required according to the default config. "Opt-in" means the dependency may be required if you change the config. And "required" just means it's always required.
            type: "opt-out"
            # An explanation of how to opt-in or opt-out. This value is optional for type=required
            how: Optionally use the selected dependency
        # Description of the dependency relationship
        description: A dependency that demonstrates the way to configure a dependent service
        # This is a list of rules that levies requirements on the configuration of the dependency and suggests ways to remedy any incompatibilities. Documentation of this feature is outside the scope of this example.
        config: ~
# This denotes any data, asset, or pointer volumes that should be connected when the "docker run" command is invoked
    # This is the image where files from the project asset directory will go
        type: data
    # This is an example of an asset volume
        type: assets
# This specifies how to configure the port mapping for exposing the service over TOR and LAN (if applicable). Many interfaces can be specified depending on the needs of the service. If it can be launched over a Local Area Network connection, specify a `lan-config`. Otherwise, at minimum, a `tor-config` must be specified.
    # This key is the internal name that the OS will use to configure the interface
        # A human readable name for display in the UI
        name: Network Interface
        # A descriptive description of what the interface does
        description: Specifies the interface to listen on for HTTP connections.
            # Port mappings are from the external port to the internal container port
                80: "80"
        # Port mappings are from the external port to the internal container port
                ssl: false
                internal: 80
        # Denotes if the service has a user interface to display
        ui: true
        # Denotes the protocol specifications used by this interface
        - tcp
        - http
# Alerts: omitting these will result in using the default alerts in EmbassyOS, except for start, which has no default.
    install-alert: This is an alert that will present before the user installs this service
    uninstall-alert: This is an alert that will present before the user uninstalls this service
    restore-alert: This is an alert that will present before the user restores this service from Embassy backup
    start-alert: This is an alert that will present before the user starts this service
# Specifies how backups should be run for this service. The default EmbassyOS provided option is to use the duplicity backup library on a system image (compat)
        # Currently, only docker actions are supported.
        type: docker
        # The docker image to use. In this case, a pre-loaded system image called compat
        image: compat
        # Required if the action uses a system image. The default value is false.
        system: true
        # The executable to run the command to begin the backup create process
        entrypoint: compat
        # Arguments to pass into the entrypoint executable. In this example, the full command run will be: `compat duplicity hello-world /mnt/backup /root/data`
          - duplicity
          - hello-world
          - /mnt/backup
          # For duplicity, the backup mount point needs to be something other than `/root`, so we default to `/root/data`
          - /root/data
            # BACKUP is the default volume that is used for backups. This is whatever backup drive is mounted to the device, or a network filesystem.
            # The value here donates where the mount point will be. The backup drive is mounted to this location.
            BACKUP: "/mnt/backup"
            main: "/root"
    # The action to execute the backup restore functionality. Details for the keys below are the same as above.
        type: docker
        image: compat
        system: true
        entrypoint: compat
          - duplicity
          - hello-world
          - /root/data
          - /mnt/backup
            BACKUP: "/mnt/backup"
            main: "/root"
# Commands that can be issued from the UI. NOTE: if no actions are required, this section can be left as an empty object
    name: Hello World Action
    description: A description that describes what the action will accomplish.
    warning: |
      A warning message indicating and potential dangers associated with the action
    # Indicates what state the service can be in while executing the action
      - running
    # Defines how the action is run
        type: docker
        image: main
        entrypoint: sh
        args: ["-c", "echo 'hello-world'"]
        # Same as note on health-checks
        inject: true
        # Required - valid values are yaml, toml, json
        io-format: json

That’s it! You now have the structure for your service’s Manifest.