Morty CLI

The Morty CLI is an interface allowing developers or users to easily interact with one or more Morty serverless instances.

Quick start

First, you need to install the Morty CLI using one of the following methods:

• Use the following command to install the latest version of the CLI : curl -fsSL https://morty-faas.github.io/install-cli.sh | sudo sh
• Download a pre-compiled binary from the releases page
• Build it from source, please see the CONTRIBUTING.md

Once you have your Morty CLI installed locally (for the rest of the commands, we assume that morty is available in your $PATH), you can list the available runtimes and choose the more appropriate for your first function :

morty runtime ls
# Or the short hand syntax
morty rt ls
# Output (can be different depending on when you execute it)
Available runtimes:
- go-1.19
- node-19
- python-3
- rust-1.67

Here we will use the node-19 runtime to create our first function :

morty function init --runtime node-19 hello-morty
# Or the shorthand syntax
morty fn init -r node-19 hello-morty

You should now have a directory hello-morty in your current directory that contains a function code example.

Once your function is ready, you will need to build it before invocating it. To do that, you need to have access to a Morty FaaS instance. Please refer to the General Documentation(TODO) to learn how to run a local Morty FaaS instance. For the sake of simplicity here, we will assume that we have a Morty Controller running on https://faas.morty.io and a Morty Registry running on https://registry.morty.io, but you can use your own values.

You need first to configure your context:

morty config add-context morty-faas --controller https://faas.morty.io --registry https://registry.morty.io
# Output
Success ! Your context 'morty-faas' has been saved and it is now the active context.

Every requests will now be performed against this Morty FaaS instance.

You can now build your function (it can take some time to complete) :

morty fn build hello-morty
# Once done, you should see something like :
Function hello-morty has been created !

Finally, you can invoke your function using the following command :

morty fn invoke hello-morty

Et voilà ! You've created your first function in Morty !

Usage

Configure your context

Morty CLI has the ability to work with multiple contexts. By default, Morty will run with the following context :

Name         : localhost
Controller URL  : http://localhost:8080
Registry URL : http://localhost:8081

To add your own context, use the following command :

morty config add-context $CONTEXT_NAME --controller=$CONTROLLER --registry=$REGISTRY

Replace $CONTEXT_NAME, $CONTROLLER, $REGISTRY with your own values.

To see all the contexts available in your configuration, use the following command:

morty config contexts

You can view the information about your current context by running the following command :

morty config current-context

By default, Morty CLI configuration is stored in ~/.morty/config.yaml. If you want, you can use a configuration file at another location by exporting the following environment variable : export MORTYCONFIG=/path/to/your/config.yaml.

To switch to a different context, run the following command :

morty config use-context $CONTEXT_NAME

Logger

A logger is embedded in the code. You need to activate it to be able to see the logs.

To enable it, simply export the following environment variable with the level you want :

export MORTY_LOG=debug

You should now see logs :

morty config current-context

# Output without MORTY_LOG set
Name         : thomas-dev
Controller URL  : http://162.38.112.57:8080
Registry URL : http://162.38.112.57:8081

# Output with MORTY_LOG
INFO[0000] Loading configuration from path: /Users/thomas/.morty/config.yaml
INFO[0000] Active context : thomas-dev
Name         : thomas-dev
Controller URL  : http://162.38.112.57:8080
Registry URL : http://162.38.112.57:8081

Create a new function

To create a new function, also called a workspace, you can use the following command:

morty function init $FUNCTION_NAME <opt:$DIRECTORY>
# or alias
morty fn init $FUNCTION_NAME <opt:$DIRECTORY>

By default, if the flag --runtime is not specified, the function will be created using the node-19 runtime.

Replace $FUNCTION_NAME with your own value. You can also specify an additional argument $DIRECTORY to control where the function will be created.

This command will create for you the skeleton of the function to enable fast development. Please note that the morty.yaml file inside the function directory is mandatory and it contains function metadata used during the build process.

Build the function

This command will send your function to the registry, and create it inside the system. It will rely on the morty.yaml file in the function directory to define function name and runtime.

morty function build $FUNCTION_DIRECTORY

Invoke a function

Once your function has been built, you can invoke it through the CLI. We assume here that you have a function hello-world already built in the system that produces a JSON output :

{
  "output": "Hello from my Hello World function"
}

To invoke your function, simply run :

morty fn invoke hello-world

# Output
{
    "output": "Hello from my Hello World function"
}

You can call your function with parameters by using the --param flag :

morty fn invoke --param name=Morty hello-world

# Output
{
    "output": "Hello Morty from my Hello World function"
}

By default, the command will send an HTTP GET request on the function endpoint to invoke it.

If you want to invoke your function with a different method, with a body or custom headers, you can use the flags of the invoke command :

For example, to send a POST request with data for your function :

morty fn invoke -X POST -d '{"foo":"bar"}' hello-world

To add custom headers to the request, for example to precise the Content-Type header, use the following command :

morty fn invoke -X POST -d '{"foo":"bar"}' -H "Content-Type: application/json" hello-world

For other methods or flags, use the command help :

morty fn invoke --help