Hi everyone, Is it possible to specify the profile you want to use in

Deployment.build_from_flow

similar to what you do when using the CLI

prefect --profile "profile1" deployment build ....

Is there a callback for when a flow starts? I see there are callbacks for on_completion and on_failure

This is regarding work queue concurrency limit, I set concurrency limit of work queue to 5, but only see 1 or 2 flowruns at the same time

is there a way to see flow runs beyond last 7 days in the deployment view?

Having trouble accessing the flow name from within a flow using the context for some reason:

Traceback (most recent call last):
  File "/home/aradox/code/wfp/wfp-prefect/utilities/slack.py", line 93, in handled_error_func
    return func(*args, **kwargs)
  File "/home/aradox/code/wfp/wfp-prefect/utilities/van.py", line 213, in alert_new_van_contact
    "flow": get_run_context().flow.name,
AttributeError: 'TaskRunContext' object has no attribute 'flow'

Did something change with how deployed flows run with the

path

argument specified? I’m using S3 storage and for each deployment I specify a custom `path`:

Deployment.build_from_flow(flow=flow, infrastructure=ECSTask.load('...'), storage=S3.load('...'), name=my_custom_name, path=my_custom_name, apply=True)

my flow assumes that the files in storage are pulled to the root during flow runs but _now it seems like it’s pulled as a subdirectory of

my__custom_name

Are there any preferred or in house strategies I should use if I’d like to version a flow somehow? For example, what if I had 2 ML models i’d like to run on my data- model 1 (v1) and model 2 v(2). Is there a way to have my flow have 2 “versions” and run either of them through the UI? For example, having a “version” parameter when running the flow which decides which version to run? Would something like this best be suited for having different “deployments” of a flow?

Hi all! I just saw the nice improvements in 2.8.7 and updated immediately. Curiously, some of the changes were UI improvements to the flow run diagram, and yet after the update the entire diagram has disappeared for me! Is there a setting a now need to toggle to bring back both that and the radar graph?

prefecthq/prefect:2.8.7-python3.11 docker container fails to start with the following error "AttributeError: module 'collections' has no attribute 'MutableMapping'". I am using the same docker-compose file I used for prefect 2.8.3 which was working fine BTW. Any leads would be really appreciated.

Is there a way to ensure each deployment can only have a single run at a time?

Hi everyone, I am trying to write functional tests for my code [not prefect] and I want to bypass the prefect decorators [task and flow] so I can write tests for my codes only. Is there any easy way of doing so WITH support of

with_options

methods?

Hi, could you please help me to specify private key when accessing SFTP? I trying to define in the settings as

key_file

but it fails with error

TypeError: connect() got an unexpected keyword argument 'key_file'

I want to create dependencies such that a flow is only run if other flows have already run successfully. I'm trying to use

run_deployment()

to create a flow of flows. But I'd like the logs to be available from the top level flow. I have tried using

log_prints=True

, but this hasn't solved the problem. Any thoughts? @James Zhang I saw you were using

run_deployment()

. Any luck surfacing logs into the parent flow?

The message from 2.8 release does a great job in explaining the work pool and work queue concepts. I feel adding the same information to this concept doc will make the doc better

how do i trigger my flow when an api is hit with parameters from that api hit

From the docs, it is still unclear to me yet • whether work queues can be created without being associated to work pools. This line in code implies work pool is optional. Is my understanding correct? • do work queue names have to be unique across all work pools. This line in code implies the work queue names have to be unique across work pools. Is my understanding correct?

A few of my flows that used to work in 2.8.6 (as well as previous versions) fail in 2.8.7 with concurrency-related errors. Should this be reported or are the changes expected to break some existing flows?

Is there a doc on Prefect's fault tolerance and recovery? I ran the following simple experiment which seems to suggest that Prefect cannot recover from server failures. So, I feel l'm missing something. The experiment: • Run the server • Run a flow (infinite loop that does nothing) in terminal -> Flow run state changes to

Running

• Kill the server • Kill the flow run in the terminal • Re-start the server -> Flow run state is still in

Running

My question is: Will the server notice that the flow has been running longer than the specified

timeout_seconds

and recover from it (have a state that will be retried)? Deployments have the same behaviour too. If my observation is correct, I can open a github issue to track this

Hi, all. We have a number of flows set up in a legacy Prefect Cloud v1 environment, all of which started failing last night at about 11pm Pacific Time with this error:

Failed to load and execute Flow's environment: HangupException('@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r\n@ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!

I can't find anything in the docs about it... anyone have any ideas where to start looking? These Flows were all working properly previously.

Hi everyone, m running a task that always fails. The task is cached and json serializer is being used. I was expecting an uncaught exception to trigger a failed task state and failed task states should not be caught. Is it expected behaviour for the json serializer to attempt serialize an exception? Ill add a sample code soon.

Hi everyone, it is my first time using Prefect and I'm not sure why the following code does not work. Basically I'm trying to set the @flow decorator to a method of a class:

from prefect import flow

class myClass:

    @flow(log_prints=True)
    def start(self):
        print("Start class")

if __name__ == "__main__":

    myclass = myClass()
    myclass.start()

This is the error that I get:

Error binding parameters for function 'start': missing a required argument: 'self'.

Thanks for your help!

Hello everyone, I'm having an issue deploying a Kubernetes Prefect 2 agent that I'm stuck on. After running a

kubectl -n <my_namespace> apply -f prefect2_agent.yaml

the deployment launches but keeps returning

Response: {'detail': 'Not Found'}

when attempting to connect to the work-queue. Even after manually exec'ing into the deployment pod and running a

prefect cloud login -k $PREFECT_API_KEY

receiving a login confirmation and choosing the correct workspace I cannot interact with any resource. For example running

prefect cloud workspace ls

raise the error

AttributeError: 'NoneType' object has no attribute 'groups'

even though the workspace was correctly set in the cloud login step. If anyone could shine a light where my issue may be I'd be grateful.

Hey Community, how's it going? Just wanted to raise this post in case more people are impacted by this error:

@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY!
Someone could be eavesdropping on you right now (man-in-the-middle attack)!
It is also possible that a host key has just been changed.
The fingerprint for the RSA key sent by the remote host is
SHA256:uNiVztksCsDhcc0u9e8BujQXVUpKZIDTMczCvj3tD2s.
Please contact your system administrator.
Add correct host key in ~/.ssh/known_hosts to get rid of this message.
Host key for <http://github.com|github.com> has changed and you have requested strict checking.
Host key verification failed.

GitHub updated their RSA SSH Host Key today. There are troubleshooting steps that are outlined in this article which should allow your flows to continue running and clear up the error. 👍

Hey, does the prefect training have limited seats?

Hola 🙂 I am using the latest Prefect version and I get the following error when I run my deployment: "NameError: name 'os' is not defined" . I think there is a bug in the "compat.py" file, actually in the copytree method (line : mode = stat.S_IMODE(st. st_mode)) I think the "os" is missing so it should be os.stat ? Greets

image.png

hey! I am trying to do prefect deployment build with github and docker blocks. I am getting a

too many values to unpack (expected 2)

error, anyone know what I might be doing wrong? im on version 2.8.6

the instructions im going off of are from sept 2022 and version 2.3.2, not sure if something has changed

What is the url to log in to prefect v1 cloud?

Not sure where to give feedback on the REST API docs at https://app.prefect.cloud/api/docs - so here goes: The initial thing I usually look for in api docs is how to pass auth in a request, and I couldn’t find that info anywhere in the docs - ended up figuring it out via searching here, but I’d think this basic info would be right at the top of the docs. In addition, it was unclear which API token would work: do service account keys work, or only personal keys? I think only personal keys work, but that’s just a guess