Commands and Script tasks

Types of tasks for executing programming languages.

For each of the supported languages (e.g., Python, R, Node.js, Shell), Kestra provides two types of tasks: Script and Commands.

The Script tasks are typically written inline in the YAML flow configuration. They are best suited for short scripts, and they make it easy to pass data from flow inputs and other tasks to your custom scripts.
The Commands tasks are best suited for more complex and longer scripts, typically integrated into Kestra as namespace files.

The following table gives an overview of all script tasks and their configuration.

Language	Default `image`	`beforeCommands` example	`Script` example	`Commands` example
Python	python	pip install requests kestra	print("Hello World!")	python hello.py
R	r-base	Rscript -e "install.packages('dplyr')"	print("Hello World!")	Rscript hello.R
Julia	julia	julia -e 'using Pkg; Pkg.add("CSV")'	println("Hello World!")	julia hello.jl
Ruby	ruby	gem install httparty	puts "Hello World!"	ruby hello.rb
Node.js	node	npm install json2csv	console.log('Hello World!');	node hello.js
Shell	ubuntu	apt-get install curl	echo "Hello World!"	./hello.bash
Powershell	powershell	Install-Module -Name ImportExcel	Write-Output "Hello World!"	.\hello.ps1
Go	golang	go mod init go_script	println("Hello World!")	go run hello.go
Deno	denoland/deno	N/A	console.log("Hello from kestra!")	deno run main.ts
Lua	nickblah/lua	N/A	print("Hello from kestra!")	lua -e 'print("Hello from kestra!")'
Bun	over/bun	bun add cowsay	console.log("Hello, World!")	bun run index.ts
PHP	php	N/A	echo "Hello, World!	php main.php
Perl	perl	N/A	print $message .	perl -e 'print "Hello from Kestra!
Groovy	groovy	N/A	- groovy --version	println "Hello, $name!"

Full class names:

Check available blueprints to get started with those tasks.

When to use `Script` over `Commands`?

The Script pattern has several advantages:

Simplicity: the script code is stored and versioned using Kestra's revision history along with your orchestration logic.
Easier integration with the templating engine: when the entire workflow is defined in one configuration file, it can be easier to access flow inputs, variable definitions, pass outputs to downstream tasks, etc.

However, the Script tasks also have some disadvantages as compared to the Commands tasks:

Readability: adding scripts into a YAML configuration file makes the scripts less readable, especially if they get long. Inline scripts also lack the syntax highlighting and testability that you would get when leveraging our embedded Code Editor instead.
Complex use cases: if your business logic comprises multiple files importing classes and functions from other modules, you need to use the Commands tasks instead.

Overall, we recommend using Commands tasks for advanced production workloads. However, the Script tasks offer a great alternative for simple use cases.

Examples

Here's an example comparing the Python Script and Commands task using the same Python Code:

Script Task:

yaml

id: python_script
namespace: company.team

tasks:
  - id: python
    type: io.kestra.plugin.scripts.python.Script
    beforeCommands:
      - pip install requests kestra
    script: |
      from kestra import Kestra
      import requests

      response = requests.get('https://kestra.io')
      print(response.status_code)

      Kestra.outputs({'status': response.status_code, 'text': response.text})

Commands Task:

yaml

id: python_commands
namespace: company.team

tasks:
  - id: python
    type: io.kestra.plugin.scripts.python.Commands
    namespaceFiles:
      enabled: true
      include:
        - main.py
    beforeCommands:
      - pip install requests kestra
    commands:
      - python main.py

main.py file:

python

from kestra import Kestra
import requests

response = requests.get('https://kestra.io')
print(response.status_code)

Kestra.outputs({'status': response.status_code, 'text': response.text})

Pass Values Into Code

You can also pass values into your code using expressions. We can put the expression directly inside of the Script task:

yaml

id: python_script_dynamic
namespace: company.team

inputs:
  - id: uri
    type: STRING
    defaults: https://kestra.io

tasks:
  - id: python
    type: io.kestra.plugin.scripts.python.Script
    beforeCommands:
      - pip install requests kestra
    script: |
      from kestra import Kestra
      import requests

      response = requests.get('{{ inputs.uri }}')
      print(response.status_code)

      Kestra.outputs({'status': response.status_code, 'text': response.text})

You can also pass values into Commands tasks by using an environment variable:

yaml

id: python_commands_dynamic
namespace: company.team

inputs:
  - id: uri
    type: STRING
    defaults: https://kestra.io

tasks:
  - id: python
    type: io.kestra.plugin.scripts.python.Commands
    namespaceFiles:
      enabled: true
      include:
        - main.py
    beforeCommands:
      - pip install requests kestra
    commands:
      - python main.py
    env:
      URI: "{{ inputs.uri }}"

main.py file:

python

from kestra import Kestra
import requests
import os

response = requests.get(os.environ['URI'])
print(response.status_code)

Kestra.outputs({'status': response.status_code, 'text': response.text})

Was this page helpful?

ScriptsProgramming Languages

ScriptsInline Scripts in Docker

​Commands and ​Script tasks

When to use Script over Commands?

Examples

Pass Values Into Code

Commands and Script tasks

When to use `Script` over `Commands`?