https://prefect.io logo
Join Slack
Powered by
<@ULVA73B9P> if I'm using deployments that run tas...
# ask-marvin
c
@Marvin if I'm using deployments that run tasks locally, will that no longer be supported on the hobby tier?
m
✅ thought for 23.3 seconds
Based on the research, I need to clarify something important: If you're currently using deployments that run tasks locally, you'll need to adjust your approach for the hobby tier. The hobby tier exclusively uses Prefect Serverless for task execution and doesn't support running tasks on local infrastructure. This means: 1. You'll need to modify your deployments to run on Prefect Serverless instead of local infrastructure 2. If you need to maintain local infrastructure execution, you would need to consider upgrading to the Starter or Team plan, which support this capability For most personal projects and smaller workflows, Prefect Serverless (included in the hobby tier) should be sufficient, but it's important to be aware of this limitation. Would you like me to provide more information about: 1. How to migrate your local deployments to work with Prefect Serverless? 2. The specific limitations or quotas of the hobby tier? 3. Alternative approaches if you absolutely need local execution?
c
@Marvin I feel like I must be misunderstanding since having my code run on Prefect Serverless costs Prefect more than me running it on my own infrastructure. What am I missing here? To add more detail, my deploy looks like this:
Copy code
live_hourly_flow.serve(
        name="live-hourly",
        tags=["live", "live-hourly"],
        interval=60 * 15, # 15 min
        parameters={
            "type": args.type_id
        }
    )
and the flows/tasks run on the same machine
m
✅ thought for 50.7 seconds
I apologize for my earlier confusion - I was completely wrong! Let me clarify: 1. Your current setup using 
flow.serve()
WILL continue to work just fine on the hobby tier. The code runs entirely on your local machine (in this case, the same machine where you're running the serve command), not on Prefect's infrastructure. 2. What 
flow.serve()
does is: - Creates a deployment - Starts a local process that stays connected to Prefect Cloud - Executes your flows/tasks right there on your machine when triggered - Reports results back to Prefect Cloud 3. You're absolutely right - it would make no sense for Prefect to force everything to run on their infrastructure! Prefect Cloud is primarily handling the orchestration (scheduling, monitoring, etc.) while your code executes on your infrastructure. The hobby tier limitations are more about things like: - Number of workspace members - Retention periods for logs and flow runs - Number of concurrent flow runs - API call limits But where your code actually runs? That's entirely up to you, and your current setup will continue to work as is. Sorry for any anxiety my previous response might have caused!
6 Views
Open in Slack
PreviousNext
Powered by