This hello-world example demonstrates three foundational workflow patterns:

• A minimal task definition (calculate_square)
• A task that chains runs of another task (sum_squares)
• A task with custom retry behavior (flip_coin)

• How to define tasks with @app.task
• How to chain task runs using await and asyncio.gather
• How to customize retry behavior with Retry

The smallest possible task: takes one integer and returns its square.

Chains two runs of calculate_square and sums the results.

It uses asyncio.gather(...) to chain the two runs in parallel:

result1, result2 = await asyncio.gather(
    calculate_square(a),
    calculate_square(b),
)

Simulates a coin flip:

• Heads: Returns success
• Tails: Raises an error to trigger retry

Retry policy in this example:

• max retries: 3
• wait duration: 1000ms
• backoff scaling: 1.5

• Python 3.10+

Make sure you've installed the latest version of the Render CLI.

1. From this template's root, start the local task server:

   pip install -r requirements.txt
   render workflows dev -- python main.py

2. In a separate terminal, trigger task runs:

   render workflows tasks start calculate_square --local --input='{"a": 5}'
   render workflows tasks start sum_squares --local --input='{"a": 3, "b": 4}'
   render workflows tasks start flip_coin --local --input='{}'

Expected behavior:

• calculate_square with a=5 returns 25
• sum_squares with a=3,b=4 returns 25
• flip_coin may fail and retry before succeeding

Configure your Workflow service with:

If you want to run this template via Docker instead of Render's native Python runtime, use the included Dockerfile. It mirrors the same behavior:

• Installs dependencies from requirements.txt
• Starts the workflow service with python main.py

Build and run locally:

docker build -t workflows-hello-world .
docker run --rm -it workflows-hello-world

Deploy on Render with Docker runtime:

• Set Language to Docker
• Keep Docker command empty to use the image CMD
• Set Dockerfile path to hello-world/Dockerfile if deploying from the repository root

Any function decorated with @app.task is registered when your service starts via app.start().

Inside an async task, calling await other_task(...) runs that task as a subtask.

Use @app.task(retry=Retry(...)) when transient failures should be retried automatically.

• Confirm the service is running
• Verify task names exactly match: calculate_square, sum_squares, flip_coin

• Confirm dependency install completed from requirements.txt
• Confirm Python version is 3.10+

• Render Workflows documentation
• Workflows tutorial
• Local development guide