HTTP interface for Foxx

Fetches a list of services installed in the current database.

Returns a list of objects with the following attributes:

• mount: the mount path of the service
• development: true if the service is running in development mode
• legacy: true if the service is running in 2.8 legacy compatibility mode
• provides: the service manifest’s provides value or an empty object

Additionally the object may contain the following attributes if they have been set on the manifest:

• name: a string identifying the service type
• version: a semver-compatible version string

Fetches detailed information for the service at the given mount path.

Returns an object with the following attributes:

• mount: the mount path of the service
• path: the local file system path of the service
• development: true if the service is running in development mode
• legacy: true if the service is running in 2.8 legacy compatibility mode
• manifest: the normalized JSON manifest of the service

Additionally the object may contain the following attributes if they have been set on the manifest:

• name: a string identifying the service type
• version: a semver-compatible version string

Installs the given new service at the given mount path.

The request body can be any of the following formats:

• application/zip: a raw zip bundle containing a service
• application/javascript: a standalone JavaScript file
• application/json: a service definition as JSON
• multipart/form-data: a service definition as a multipart form

A service definition is an object or form with the following properties or fields:

• configuration: a JSON object describing configuration values
• dependencies: a JSON object describing dependency settings
• source: a fully qualified URL or an absolute path on the server’s file system

When using multipart data, the source field can also alternatively be a file field containing either a zip bundle or a standalone JavaScript file.

When using a standalone JavaScript file the given file will be executed to define our service’s HTTP endpoints. It is the same which would be defined in the field main of the service manifest.

If source is a URL, the URL must be reachable from the server. If source is a file system path, the path will be resolved on the server. In either case the path or URL is expected to resolve to a zip bundle, JavaScript file or (in case of a file system path) directory.

Note that when using file system paths in a cluster with multiple Coordinators the file system path must resolve to equivalent files on every Coordinator.

Removes the service at the given mount path from the database and file system.

Returns an empty response on success.

Removes the service at the given mount path from the database and file system. Then installs the given new service at the same mount path.

This is a slightly safer equivalent to performing an uninstall of the old service followed by installing the new service. The new service’s main and script files (if any) will be checked for basic syntax errors before the old service is removed.

The request body can be any of the following formats:

• application/zip: a raw zip bundle containing a service
• application/javascript: a standalone JavaScript file
• application/json: a service definition as JSON
• multipart/form-data: a service definition as a multipart form

A service definition is an object or form with the following properties or fields:

• configuration: a JSON object describing configuration values
• dependencies: a JSON object describing dependency settings
• source: a fully qualified URL or an absolute path on the server’s file system

When using multipart data, the source field can also alternatively be a file field containing either a zip bundle or a standalone JavaScript file.

When using a standalone JavaScript file the given file will be executed to define our service’s HTTP endpoints. It is the same which would be defined in the field main of the service manifest.

If source is a URL, the URL must be reachable from the server. If source is a file system path, the path will be resolved on the server. In either case the path or URL is expected to resolve to a zip bundle, JavaScript file or (in case of a file system path) directory.

Note that when using file system paths in a cluster with multiple Coordinators the file system path must resolve to equivalent files on every Coordinator.

Installs the given new service on top of the service currently installed at the given mount path. This is only recommended for switching between different versions of the same service.

Unlike replacing a service, upgrading a service retains the old service’s configuration and dependencies (if any) and should therefore only be used to migrate an existing service to a newer or equivalent service.

The request body can be any of the following formats:

• application/zip: a raw zip bundle containing a service
• application/javascript: a standalone JavaScript file
• application/json: a service definition as JSON
• multipart/form-data: a service definition as a multipart form

A service definition is an object or form with the following properties or fields:

• configuration: a JSON object describing configuration values
• dependencies: a JSON object describing dependency settings
• source: a fully qualified URL or an absolute path on the server’s file system

When using multipart data, the source field can also alternatively be a file field containing either a zip bundle or a standalone JavaScript file.

When using a standalone JavaScript file the given file will be executed to define our service’s HTTP endpoints. It is the same which would be defined in the field main of the service manifest.

If source is a URL, the URL must be reachable from the server. If source is a file system path, the path will be resolved on the server. In either case the path or URL is expected to resolve to a zip bundle, JavaScript file or (in case of a file system path) directory.

Note that when using file system paths in a cluster with multiple Coordinators the file system path must resolve to equivalent files on every Coordinator.

Fetches the current configuration for the service at the given mount path.

Returns an object mapping the configuration option names to their definitions including a human-friendly title and the current value (if any).

Replaces the given service’s configuration.

Returns an object mapping all configuration option names to their new values.

Replaces the given service’s configuration completely.

Returns an object mapping all configuration option names to their new values.

Fetches the current dependencies for service at the given mount path.

Returns an object mapping the dependency names to their definitions including a human-friendly title and the current mount path (if any).

Replaces the given service’s dependencies.

Returns an object mapping all dependency names to their new mount paths.

Replaces the given service’s dependencies completely.

Returns an object mapping all dependency names to their new mount paths.

Fetches a list of the scripts defined by the service.

Returns an object mapping the raw script names to human-friendly names.

Runs the given script for the service at the given mount path.

Returns the exports of the script, if any.

Runs the tests for the service at the given mount path and returns the results.

Supported test reporters are:

• default: a simple list of test cases
• suite: an object of test cases nested in suites
• stream: a raw stream of test results
• xunit: an XUnit/JUnit compatible structure
• tap: a raw TAP compatible stream

The Accept request header can be used to further control the response format:

When using the stream reporter application/x-ldjson will result in the response body being formatted as a newline-delimited JSON stream.

When using the tap reporter text/plain or text/* will result in the response body being formatted as a plain text TAP report.

When using the xunit reporter application/xml or text/xml will result in the response body being formatted as XML instead of JSONML.

Otherwise the response body will be formatted as non-prettyprinted JSON.

Puts the service into development mode.

While the service is running in development mode the service will be reloaded from the filesystem and its setup script (if any) will be re-executed every time the service handles a request.

When running ArangoDB in a cluster with multiple Coordinators note that changes to the filesystem on one Coordinator will not be reflected across the other Coordinators. This means you should treat your Coordinators as inconsistent as long as any service is running in development mode.

Puts the service at the given mount path into production mode.

When running ArangoDB in a cluster with multiple Coordinators this will replace the service on all other Coordinators with the version on this Coordinator.

Fetches the Swagger API description for the service at the given mount path.

The response body will be an OpenAPI 2.0 compatible JSON description of the service API.

Downloads a zip bundle of the service directory.

When development mode is enabled, this always creates a new bundle.

Otherwise the bundle will represent the version of a service that is installed on that ArangoDB instance.

Commits the local service state of the Coordinator to the database.

This can be used to resolve service conflicts between Coordinators that cannot be fixed automatically due to missing data.