Temporal Server

  • Commercial Systems
Channel Revision Published Runs on
latest/stable 23 08 Mar 2024
Ubuntu 22.04
latest/edge 42 14 Nov 2024
Ubuntu 22.04
juju deploy temporal-k8s --channel edge
Show information

Platform:

Run Your First Workflow

This is part of the Charmed Temporal Tutorial. Please refer to this page for more information and the overview of the content.

Now that we have a Temporal Server up and running with a Temporal Worker connected to it, we can trigger a workflow execution and observe its result in the Web UI.

To do so, we can create the following simple client, which triggers the named workflow GreetingWorkflow defined in the worker.

from temporallib.client import Client, Options
import asyncio

async def main():
    client_opt = Options(
        host="10.1.232.64:7233", # Replace with your Temporal Server unit IP address
        queue="test-queue",
        namespace="default",
    )

    client = await Client.connect(client_opt=client_opt)
    workflow_name = "GreetingWorkflow"
    workflow_id = "hello-activity-workflow-id"

    greeting = await client.execute_workflow(
        workflow_name,
        "World",
        id=workflow_id,
        task_queue="test-queue",
    )

    logger.info(f"Greeting: {greeting}")


if __name__ == "__main__":
    asyncio.run(main())

Install the necessary packages and run the script above using:

pip install temporal-lib-py

python workflow.py

# Output:
Greeting: Hello, World!

Further details on the workflow execution can be viewed on the web UI.

Note: Workflow and activity payloads can be encrypted by setting the same key in the encryption-key config of the Temporal worker and the client_opt in the code above.

If using a TLS connection through ingress, you must provide the certificate generated in the ingress step of this tutorial as part of your TLSConfig when making requests to the Temporal server.

See next: Cleanup your environment


Help improve this document in the forum (guidelines). Last updated 1 year, 2 months ago.