Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

Light python framework for AWS SWF

Project description

Light python framework for AWS SWF

About this Project

Caravan is being used at Ludia for projects like marketing campaign system and on-demand distributed processing systems (thanks to hundreds of Lambda functions).

Feedbacks, ideas and contributions are highly welcomed. (Just open a Github issue).

Focus of this project

Similar projects exists (like Simpleflow which seems mature). Here is where Caravan differs from existing projects:

  • Support of AWS Lambda tasks
  • Boto3
  • KISS
  • Bring your own workflow framework (standard implementations as contribs)
  • No coupling between Decider code and Activity code
  • Paster compatible config file


  • Decider worker
  • Activity task worker TODO
  • Commands to start/signal/terminate an arbitrary workflow execution
  • Command to list open workflow execution
  • Command to register a domain / list domains


Caravan uses Boto3 to connect to AWS. See Boto 3 configuration guide for the complete documentation.


On AWS EC2, the metadata provider only provides the credentials, the must be provided by configuration.

Environment Variables

The access key for your AWS account.
The secret key for your AWS account.
The default region to use, e.g. us-east-1.
The default credential and configuration profile to use, if any.

Configuration Files

The credentials file is located at ~/.aws/credentials:

# The access key for your AWS account
aws_access_key_id=<YOUR ACCESS KEY ID>

# The secret key for your AWS account
aws_secret_access_key=<YOUR SECRET KEY>

The settings file is located at ~/.aws/config:

# The default region when making requests
region=<REGION NAME>

It also supports profiles:

[profile dev-profile]
# The default region when using the dev-profile account
region=<REGION NAME>


Setup a SWF domain to run this example:

$ caravan-domain-register -n CaravanDemo --retention-days 1

Write a workflow type (see full demo)

from caravan import Workflow

class Demo(Workflow):

    """Noop workflow using the bare caravan API."""

    name = 'Demo'
    version = '0.1'
    default_execution_start_to_close_timeout = '600'
    default_task_start_to_close_timeout = '10'

    def run(self):

Run the decider with the Demo workflow:

$ caravan-decider -d CaravanDemo -m caravan.examples.demo -t default --verbose

Start an execution of the Demo workflow:

$ caravan-start -d CaravanDemo -n Demo -v 0.1 -i 1

(The Demo workflow will wait for 5 minutes)

List the executions:

$ caravan-list -d CaravanDemo
$ caravan-list -d CaravanDemo --oldest 2015-01-01

Send a signal to an execution:

$ caravan-signal -d CaravanDemo -i 1 -s PRINT --input 'Hello World!'
$ caravan-signal -d CaravanDemo -i 1 -s PRINT --input 'Lorem ipsum'
$ caravan-signal -d CaravanDemo -i 1 -s STOP

Terminate an execution:

$ caravan-terminate -d CaravanDemo -i 1


Possibly use virtualenvwrapper to manage your virtualenvs.

$ mkvirtualenv caravan


$ virtualenv caravan

Clone and install development dependencies:

(caravan)$ git clone
(caravan)$ cd caravan
(caravan)$ pip install -e .[dev]

Run tests:

(caravan)$ nosetests


The release process use zest.releaser:

$ fullrelease


MIT licensed. See the bundled LICENSE file for more details

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for caravan, version 0.2.2
Filename, size File type Python version Upload date Hashes
Filename, size caravan-0.2.2-py2-none-any.whl (32.1 kB) File type Wheel Python version py2 Upload date Hashes View hashes
Filename, size caravan-0.2.2.tar.gz (19.7 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page