Saga orchestration with Catalyst Workflow

This repo contains a Python app that shows you how to implement the AWS reference architecture for Saga orchestration using Catalyst Workflow.

Catalyst Workflow is a fully managed durable exection engine built on top of the open-source Dapr project. With Catalyst Workflow on AWS you can implement the saga pattern via orchestration and get the following benefits:

• Use any AWS compute service available, including ECS, AppRunner, EC2 etc.
• Use a fully managed Diagrid state store for workflow execution state, or bring your own
• Use any messaging service for downstream notifications, from Diagrid's managed pub/sub to your own Kafka, SQS etc.
• End to end observability with API logs, metrics and more
• A workflow execution UI that allows you to dig into each activity's execution status
• A single and consistent polyglot programming model that runs anywhere
• Local debugging without needing emulators or Docker containers

Note: You can run Catalyst Workflow on any cloud or compute environment including your local machine

Running the sample

Prerequisites

The following services, tools, and frameworks are required for this demo:

• Diagrid Catalyst account (sign up for private beta access) and the Diagrid CLI
• Python 3.8 or later

Getting started

Create a new Catalyst project named catalyst-workflow and use the Diagrid managed pub/sub broker & KV store, and enable the managed workflow API:

diagrid project create catalyst-workflow --deploy-managed-pubsub --deploy-managed-kv --enable-managed-workflow --wait

To set this project as the default in the CLI run:

diagrid project use catalyst-workflow

Create a new App ID for our reservations app:

diagrid appid create reservations

In the root of the Python app, set the project URLs and API token for Catalyst:

export DAPR_HTTP_ENDPOINT=$(diagrid project get -o json | jq -r '.status.endpoints.http.url')
export DAPR_GRPC_ENDPOINT=$(diagrid project get -o json | jq -r '.status.endpoints.http.url')
export DAPR_API_TOKEN=$(diagrid appid get reservations -o json | jq -r '.status.apiToken')

All done! Now you can run the app, issue a JSON request and see the workflow run to completion:

python3 app.py

curl -XPOST -H "Content-type: application/json" -d '{"first_name": "hello", "last_name": "there", "location": "spain", "car_class": "compact"}' 'http://localhost:5000/reserveFlight'

Running in a container

You can containerize this code and run it on any container based environment:

docker built -t <REPO>/<NAME>:latest .

Note: make sure to supply the container with the environment variables for the http, grpc endpoints and the API token

Bringing your own database and messaging services

This example uses the Catalyst fully managed database and pub/sub brokers for ease of use and easy set up experience. If you want to use your own cloud services, read here.