Hi, Community, I tried to follow the tutorial of Prefect Cloud 2.0. I could run the flow locally and saw the flow run results right there in Prefect Cloud. However, when I created a flow run from my deployment on the cloud, the local terminal session didn't pick up the flow run and kept showing the error below. Any help will be appreciated!

_from_ prefect.utilities.asyncio _import_ Sync

hi everyone, I wonder if what is the equivalent thing for this in prefect 2.0 ? I'm trying to implement something similar to

as_completed()

from Dask, thank you.

Hi everyone, super specific question maybe, but at the same time I feel like we’re probably not the only one in this boat. Firstly, congrats on the release! Woke up this morning and now we can celebrate as a team that Prefect Orion that we have been using will now be stable, for which the timing is quite perfect for our team as well since we have not really gone into production yet. With that being said, we stayed with Prefect 2.0b8 since we did not want to update every day of the week since last week or so, but now we feel a little lost in all the changes with I think the last version technically being 2.0b16. A lot of the announcements/tutorials etc. are focused on migration from prefect 1.0 (understandably), and I expect significant changes in the docs since Prefect 2.0b8. However, is there a way to somehow easily see differences in the docs since for example 2.0b8 (or earlier releases), or do we basically have to read through the entire docs again? I am already glad I realized the release notes are still in this file: https://github.com/PrefectHQ/prefect/blob/2.0.0/RELEASE-NOTES.md because previous times this file was linked but this time it was not 😛

Hey prefect team, Great job on getting to 2.0 GA !!!! I had question about the new deployment method is there example where I can easily add a customizations to a K8 job without going into the ui and creating k8 job block? Previously all i had to do was import the k8 job create a deployment and pass my customizations to it.

Hey, I've encountered an issue or misunderstanding with the prefect 2.0.0 manifests or deployment packages: I've configured a a GCS storage block called gcs-prefect-stprage and used this in the deployment build step. My folder structure is as follows:

test
  |___ deployment.ymal
  |___ kayak_database_flows.py

My build command is (run from inside the folder "test")

prefect deployment build kayak_database_flows.py:integrate_kayak_details --name integrate_kayak_db_details -t rmesbtest1 -sb gcs/gcs-prefect-stprage

Then i apply the deployment and run the deployment. Everything fine so far, the flow get's picked up by the corresponding agent However, the agent downloads ALL of the content of the GCS bucket. So really all of the content in this bucket, not only the stuff related to this deployment. Is this expected? If so, do I need a separate GCS storage block for any of the deployments?

You guys should put up some more documentation regarding the vertex agent. It is absolutely a fantastic way to run pipelines. Am currently running the vertex agent in a GKE autopilot cluster. The vertex ai pipelines really allows you to scale endlessly while at the same time massively reducing costs.

Hi. Quick question on flow deployments - I notice if I rename a flow, the previous named flow still exists and the newly named one is created as a new separate flow. Is there any way to unregister/delete the old flow) programmatically?

Hi all and congrats on the new GA release! I have a question: In latest beta versions we were able to have multiple Deployments specified in a yaml file separated with --- and then by running

prefect deployment create deployments.yaml

we were able to deploy them all at once in Prefect Server. I have seen that there are a lot of changes in GA compared to latest beta and I took a look at the updated docs about it. However I think that some (important to me) functionality is missing compared to what we had in beta. Is there an easy way in the new GA deployments procedure to have multiple flows in a yaml file or an alternative way to deploy multiple flows at once without having to call

prefect deployment build ......

and

prefect deployment apply deployment.yaml

manually for each single flow that we have?

Hello Prefect, short question: When I register my flow I use

labels='prod'

however the registered flow also uses the id of the docker instance it is built into (See screenshot): Is there anyway to delete that, so that only the labels I give are used?

issue with my api_url prefect.exceptions.PrefectHTTPStatusError: Client error '403 Forbidden' for url 'https://api-beta.prefect.io/api/accounts/xxxxxxxxxxxxxx/workspaces/xxxxxxxxxxxxxxxxxx/block_types/name/Process' Response: {'detail': 'Not authenticated'} For more information check: https://httpstatuses.com/403 ps : changed my account id and workspace id to xxxxx to post the issue here

Hi, I have a question about why task args cannot have copy, this seems affect my creat flow run properly using trigger arg See thread for code

Hi All, I am using prefect server deployed on K8. For creating environment when prefect flows are deployed on k8 pods,I am trying to use conda environment. what command does prefect agent run while starting a pod ? As we are getting

exit error 2, Instance terminated

on dashboard and on pod logs we are getting

No such command "execute"

hi all, I’ve upgraded to Prefect 2.0.0, now how can I define a block for Azure blob storage? I just see S3 and GCP blocks…

Good morning, quick questions 1. can one agent run two different flows in parallel? 2. can 2 different agents run 2 different flows at the same time? 3. is there a limit on the number of agents we can start?

Are there any examples of adding loggers to prefect for prefect 2.0?

Hello thanks for the hard work prefect 2.0 seems really promising 👏 I did manage to work a very simple hello world in a kubernetes job and have just a question, is that still possible to have the ouptut of a python function in the job run log ? From what i have seen i can’t see the ouptput (picture in thread)

👋 working on getting prefect 2.0 setup for our team, and still learning how to navigate the new docs. Is there a way to set the current flow run to cancelled from within a task or flow?

Hi thank you for the amazing job on releasing 2.0. I am trying the new deployment system with a

local file system

storage/block: my-local-storage. However when I try to build with the parameter "prefect deployment build --storage-block local-file-system/my-local-storage" I get an exception :

File "/usr/local/lib/python3.10/site-packages/prefect/filesystems.py", line 114, in put_directory
    shutil.copytree(from_path, local_path, dirs_exist_ok=True)
NameError: name 'from_path' is not defined

Looking at the code : https://github.com/PrefectHQ/prefect/blob/2.0.0/src/prefect/filesystems.py#L114 it seems the variable "from_path" is indeed not defined.

Where is the documentation for

prefect_test_harness

located? It tries to connect to a server, I want to skip that step

meaning, I want to ignore the

@flow

decorator.

I’ve noticed that in Prefect 2 cloud UI, when creating a KubernetesJob, it is not possible to set

image_pull_policy

, is it going to be added in the upcoming releases?

https://prefect-community.slack.com/archives/CKNSX5WG3/p1659017220160889 Shoutout to our keyboard winners from our DC Meetup yesterday! You can snag your own keyboard by writing a Prefect 2.0 flow and posting it to #show-us-what-you-got. Looking forward to seeing the funniest/weirdest/most silly flows you have to offer!

I have a parameter X and a StartFlowRun task whose flow_name is mapped over the parameter X. When I run this flow, the generated task runs of StartFlowRun don’t have much information about what value of X they received. What’s the best solution here? I don’t see how we can map over the task run name, since that is set at init time.

Hi, Congrats on the new release! Looking forward to what the future will bring! 🚀 I was wondering what the preferred route is for stopping a flow while it's running? Will there be an option to stop a continuous flow (that spins up tasks continuously) and restart it?

this may be a stupid question, but I’ve tried to build a deployment with:

prefect deployment build flows/hello_world.py:hello_world --name "Hello world dev" --tag kubernetes --tag dev --infra kubernetes-job --infra-block base-dev

and I got this:

File "/Users/tessellir/Projects/il-da-orchestrator/.venv/lib/python3.8/site-packages/prefect/blocks/core.py", line 501, in load
    block_type_slug, block_document_name = name.split("/", 1)
ValueError: not enough values to unpack (expected 2, got 1)
An exception occurred.

I’ve ofc created a block in Prefect cloud of type KubernetesJob called base-dev, what am I missing?

Hi all, this is a multipart question. For context, we are building a large Flow that has multiple components that are dependent on one another. For example we have 5 Flows A B C D E. D is dependent on A+B, E on D etc... Certain Flows are also dependent on external data sources from databases or APIs. We are implementing a retry system and a notifications system for these Flows. 1. Is there a way to limit the slack notification to only send on the LAST retry attempt if it fails? 2. For retries, is it possible to set up a retry for only specific errors? For example, if a query times out/connection fails we retry else if there is a syntax error in the query don not retry. 3. Can we send a message for certain error for example an unexpected empty dataframe but still for the flow to complete? 4. Can we trigger a retry of a flow from a different starting point? Let's say there is a failure in the 4th task but we want to retry the run from task 2? Thank you for any insights on any of these!

Is there an expected EOL for Prefect 1.0 at this time?

Is it still possible to package flows in Docker images? I don't immediately see such an option, and can't seem to figure it out

Hello. In Prefect 1, is there an easy way to make a task ONLY dependent on the exactly 1 upstream task, not all upstream ones? What I mean is

A -> B (trigger: A finished) -> C (trigger: B successful)

I want

C

to ignore what happens to

A

, but only run if

B

is successful.

B

will run after

A

is finished (no matter fail or success) I looked into state and state_handlers, seemed too complicated for such a (seemingly) simple task. edit: In other words, I don't want the trigger function to aggregate ALL the upstream tasks. Just consider the task that I've explicitly set as upstream Thanks