Skip to content

Orchestrator-Core

Production ready Orchestration Framework to manage product lifecycle and workflows. Easy to use, Built on top of FastAPI

Downloads Coverage Pypi



The orchestrator core is a project developed by SURF to facilitate the orchestration of services. Together with ESnet this project has been opensourced in the commons conservancy to help facilitate collaboration. We invite all who are interested to take a look and to contribute!

Orchestration

When do you orchestrate and when do you automate? The answer is you probably need both. Automation helps you execute repetitive tasks reliably and easily, Orchestration adds a layer and allows you to add more intelligence to the tasks you need to automate and to have a complete audit log of changes.

Orchestrate* - Transitive Verb

/ˈôrkəˌstrāt/ /ˈɔrkəˌstreɪt/

1: Arrange or score (music) for orchestral performance.

‘the song cycle was stunningly arranged and orchestrated’

2: Arrange or direct the elements of (a situation) to produce a desired effect, especially surreptitiously.

‘the developers were able to orchestrate a favorable media campaign’

Project Goal

This orchestrator-core provides a framework through which you can manage service orchestration for your end users. It enables you to define products to which users can subscribe, and helps you intelligently manage the lifecycle of Creation, Modification, Termination and Validation of a customer subscription.

Usage

This project can be installed as follows:

Step 1:

Install the core. 

pip install orchestrator-core

Step 2:

Create a postgres database: 

createuser -sP nwa
createdb orchestrator-core -O nwa

Step 3 (optional):

When using multiple workers, you will need a redis server for live updates with websockets.

By default it will use memory which works with only one worker. 

export WEBSOCKET_BROADCASTER_URL="memory://"

For the redis connection you need to set the env variable with the connection url. 

export WEBSOCKET_BROADCASTER_URL="redis://localhost:6379"

Websockets can also be turned off with: 

export ENABLE_WEBSOCKETS=False

More broadcaster info here

If you want to use pickle for CACHE serialization you will need to set the CACHE_HMAC_SECRET: 

export CACHE_HMAC_SECRET="SOMESECRET"
NOTE: The key can be any length. However, the recommended size is 1024 bits.

Step 4:

Create a main.py file.

from orchestrator import OrchestratorCore
from orchestrator.cli.main import app as core_cli
from orchestrator.settings import AppSettings

app = OrchestratorCore(base_settings=AppSettings())

if __name__ == "__main__":
    core_cli()

Step 5 (Optional):

OrchestratorCore comes with a graphql interface that can to be registered after you create your OrchestratorApp. If you add it after registering your SUBSCRIPTION_MODEL_REGISTRY it will automatically create graphql types for them. More info can be found in docs/architecture/application/graphql.md

example: 

from orchestrator import OrchestratorCore
from orchestrator.settings import AppSettings
app = OrchestratorCore(base_settings=AppSettings())
# register SUBSCRIPTION_MODEL_REGISTRY
app.register_graphql()

Step 6:

Initialize the migration environment. 

PYTHONPATH=. python main.py db init
PYTHONPATH=. python main.py db upgrade heads

Step 7:

Profit :)

Authentication and authorization are default enabled, to disable set OAUTH2_ACTIVE and OAUTH2_AUTHORIZATION_ACTIVE to False.

uvicorn --reload --host 127.0.0.1 --port 8080 main:app

Visit http://127.0.0.1:8080/api/redoc to view the api documentation.

Setting up a development environment

To add features to the repository follow the following procedure to setup a working development environment.

Installation (Development standalone)

Install the project and its dependencies to develop on the code.

Step 1 - install flit:

python3 -m venv venv
source venv/bin/activate
pip install flit

Step 2 - install the development code:

Danger

Make sure to use the flit binary that is installed in your environment. You can check the correct path by running 

which flit

To be sure that the packages will be installed against the correct venv you can also prepend the python interpreter that you want to use:

flit install --deps develop --symlink --python venv/bin/python

Running tests

Run the unit-test suite to verify a correct setup.

Step 1 - Create a database

createuser -sP nwa
createdb orchestrator-core-test -O nwa

Step 2 - Run tests

pytest test/unit_tests
or with xdist:

pytest -n auto test/unit_tests

If you do not encounter any failures in the test, you should be able to develop features in the orchestrator-core.

Installation (Development symlinked into orchestrator SURF)

If you are working on a project that already uses the orchestrator-core and you want to test your new core features against it, you can use some flit magic to symlink the dev version of the core to your project. It will automatically replace the pypi dep with a symlink to the development version of the core and update/downgrade all required packages in your own orchestrator project.

Step 1 - install flit:

python - m venv venv
source venv/bin/activate
pip install flit

flit install --deps develop --symlink --python /path/to/a/orchestrator-project/venv/bin/python

So if you have the core and your own orchestrator project repo in the same folder and the main project folder is orchestrator and you want to use relative links, this will be last step:

flit install --deps develop --symlink --python ../orchestrator/venv/bin/python

Increasing the version number for a (pre) release.

When your PR is accepted you will get a version number.

You can do the necessary change with a clean, e.g. every change committed, branch:

bumpversion patch --new-version 0.4.1-rc3