Hi all, my company has been using prefect using a cloud run work pool for nearly 1 year now. I was investigating push work pools, so I deleted my old cloud run work pool temporarily (since free tier accounts have a max of 1 allowed). Then when I went to create a new work pool I am only able to use prefect’s managed work pool without upgrading?! The managed work pool doesn’t even allow for custom docker images. I was not clear that deleting my old cloud run work pool would regress the amount of functionality I could have on my free tier account! Please could someone advise what my options are here? Besides upgrading to pay 100$ / month?

Has anyone been having issues with pulls_from_block on github? Since yesterdays github outage i’ve been having

Failed to pull from remote:
 Cloning into '/tmp/tmp0h4b39iiprefect'...
fatal: unable to access '<https://github.com/org/repo.git/>': Failed to connect to <http://github.com|github.com> port 443 after 131017 ms: Couldn't connect to server

errors on most of my deployments

Hi all, has anyone been having issues with prefect cloud automation? Usually during midnight, it fails to report/detect the failed deployments/flows. Parent flow and child flow failed at the same timestamp. Parent flows have

prod

tag but child flows not. anyone with the same issue? during daytime its fine. super weird.

Hi, I just signed up for a hobby tier Prefect plan to try things out but can't seem to create even one workspace (which I think is required for me to do anything). What am I missing?

Hi all, why can't I add my github credentials to my prefect block? Every time I try to, it just says None in the github repository block??

Is the

prefecthq/prefect:3-latest

docker image using the latest UX? It looks different from the prefect cloud UX, but I could be wrong

Hi! I want to move my pool worker to Kubernetes, but it seems the free tier doesn’t allow adding a worker pool. I started using Prefect two years ago and it was free back then. If I delete my current worker pool, will I be unable to create a new one?

Hi Team - I signed up to evaluate perfect with a local worker. Didn't do anything, except to deploy one sample deployment and now it says "Cannot cancel subscription" . Everything is under the limit of starter pack. Subscription is a subscription and should be able to cancel and not forcing anything for a customer , right?

Morning, I've contacted support already, but is anyone else who uses cloud run workers seeing infrastructure errors due to prefect creating 3-4

CreateJob

requests in a row? Results in a 409 error for us, appears like so in logs:

Hello! the service started deteriorating in Prefect Cloud, for instance, we woke up to 10 crashed flows because what seems to be a rate limiting rule

Failed to generate job configuration: Client error '429 Too Many Requests' for url '<http://orion-internal:4200/api/accounts>...
For more information check: <https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/429>

a few more also were stuck at the pending state.. is this sort of rate limiting documented somewhere for the Pro Plan? is it something we can control?

Can 'Secret' block types still be accessed in 3.6.2? I can see the secrets on the website ux, but 'prefect block ls' and python code using Secret.load('[Secret_name]') can't seem to find secrets.

Hi ? How do we set minimum resources for a prefect deployment ? I have kubernetes workers, and I need to tell them how much memory my script need to be able to autoscale my kubernetes job when deployment is running... (in Prefect 3x btw). ChatGPT give me this but I dont see any doc about it :

deployments:
- name: my_deployment
  work_pool:
    name: my_worker_pool
    job_variables:
      pod:
        spec:
          containers:
            - name: flow
              resources:
                requests:
                  cpu: "2"
                  memory: "8Gi"
                limits:
                  cpu: "4"
                  memory: "8Gi"

Hi, would someone be able to review https://github.com/PrefectHQ/prefect/pull/19498 ? Thanks!

Hey all, we're running into this scheduling issue, breaking our production pipelines. Is there an ETA on a fix here? Folks are starting to push for moving off of Prefect due to this.

hey everyone, started getting this weird issue with my Cloud Run Jobs based work pool recently. I have a flow that submits 7 flows (they are actually standalone flows with their own infra, not actually subflows). Each flow trigger its own cloud run job with a unique (prefect generated) name etc. Still I'm getting this GCP error claiming resource already exists. It seems to happen quite randomly. As you can see some of them launch successfully and sometimes all of them do. However as you can imagine it's a big problem when some of them don't. Anyone has an idea of what it can be? When I look at Cloud run I never see the job existing at any time that would justify a duplication. However it's also probably because we have

keep_jobs

set to false so even if it's duplicating some job it's probably also deleting. We're not tracking the flow state or anything like that from python, we just "fire and forget" these flows. This is my code to launch them:

for model in collection_models:
        scenarios = model_to_scenarios.get(model, [])
        if not scenarios:
            logger.warning("No scenarios found for model %s, skipping", model)
            continue

        # Submit task for this model's subflow with model-specific name
        future = launch_model_subflow_task.with_options(
            name=f"launch-{model}",
            retries=2,
            retry_delay_seconds=60,
        ).submit(
            model=model,
            scenarios=scenarios,
            collection_limit_per_brand_model=request.collection_limit_per_brand_model,
        )
        deployment_futures.append(future)

and the definition:

flow_run = await run_deployment(  # type: ignore[misc]
            name=f"process-batch-subflow/{PREFECT_DEPLOYMENT_SUFFIX}-subflow",
            parameters={
                "model": model,
                "collection_scenarios": scenarios,
                "collection_limit_per_brand_model": collection_limit_per_brand_model,
            },
            job_variables={
                "cpu": CPU_CORES,
                "memory": "32G",
            },
            timeout=0,  # Fire-and-forget: don't wait for subflow completion (execution timeout is set on the subflow itself)
            tags=[model],
        )

@Nate, any chance you can approve the auto-bot updates for conda

pydocket-feedstock

to unblock conda

prefect-feedstock

? This will help us update our env, which would be nice! https://github.com/conda-forge/pydocket-feedstock/pulls

hi all, question about the

ProcessPoolTaskExecutor

(or really, any of the parallel task runners). Since tasks can be nested inside of tasks, what's the execution model when i

task.submit()

a task, and it calls

another_task()

inside of it, where does

another_task

run? in the process that is handling

task

? i ask because it used to be that we needed to wrap tasks in flow runs to control which task runner to use

Hey all, we're seeing weird behavior with Prefect Cloud jobs when a single flow run is picked up by two ECS tasks and both execute the flow simultaneously. The logs in Prefect Cloud then appear duplicated, and one of the ECS tasks eventually fails with exception

prefect.exceptions.MissingResult: The result was not persisted and is no longer available

We observed this issue three times, first one on Saturday, then Monday, and today. This is the first time we encountered this issue after using Prefect Cloud for over 2 years now. We're using ECS Push work pools. Does anyone experience the same issue?

Hello! I'm self-hosting Prefect with SQLite and I am getting a lot of these errors: https://gist.github.com/valleybay/bb1bd9717596826e125b0343fb20c64a. It says «database is locked». It's repeating every few minutes. Anyone else experiencing this? Have you been able to get Prefect to run smoothly on SQLite?

Hi @Marvin I am hosting Prefect OSS v3.6.3 on kubernetes. I have an automation that send a slack alert when a flow enters TimedOut, Crashed, or Failed. Yesterday I got some alerts in slack, it was working. Now when a flow crashes, nothing arrives on slack even though the automation is turned on. The slack block is working: when I run these lines locally, I get the alert:

>>> from prefect.blocks.notifications import SlackWebhook
>>> slack_webhook_block = SlackWebhook.load("slack-prefect-prod-failures-v2")
>>> slack_webhook_block.notify("Hello from Prefect!")

I create the automation locally with

prefect automation create --from-file automations.yaml

. Creating the simplest automation in the UI doesn't solve the problem. What could be the cause ? I did not changed much things since yesterday.

Hi. What might cause the Prefect UI to not display task logs in the task log view, and instead display:

This run didn't generate Logs

Task logs exist in the DB and do display correctly in the flow log view. The task logs used to display at the task log view until about 2 days ago. It may have something to do with a single new flow which was added which called

get_run_logger

, a pattern which did not exist previously, has since been removed, but the problem remains globally for all our flows.

Trying to get this PR across the line: https://github.com/PrefectHQ/prefect/actions/runs/19767933876/job/56645184779?pr=19559 But I'm seeing CI fail on python3.10, python3.12 (passes for 11, 13, 14). I suspect flakiness could be a problem since there shouldn't really be differences exposed by this change in terms of python version (just adds a setting). Is this a known problem with the Github CI suite for Prefect?

Hi @Marvin, how to use environment variables placeholder in prefect.toml?

@Marvin What is the difference between call task.delay() and wrap a task into a flow? My senario is invoking long running logic through fastapi, thus I want to query the task state, task log, task result etc through my frontend.

@Marvin What is the difference between "calling a task directly in a flow" and "calling task.submit() within a flow"?

Hello everyone, I am running Prefect on my local machine. I am using

task.submit()

followed by

task.result()

to retrieve the output. However, even a simple Python function (e.g., retrieving a value from a dictionary) takes around 30 seconds to complete.

#CL09KU1K7 #C04DZJC94DC

Hi folks 👋 on my on-prem setup with kubernetes work pool, we’re trying to use OpenTelemetry in the flow runs, sinking the telemetry data to a Otel Collector, but we’ve observed two things:

OTEL_PYTHON_LOGGING_AUTO_INSTRUMENTATION_ENABLED=true

does not work with formatting the log output. Adjusting

PREFECT_FORMATTERS_*

leads to missing otel’s injected properties. traces and metrics set inside the flow code does not propagate to the Otel collector is there a magic trick to have Otel working with Prefect?

Hello in my AWS worker I am seeing an error: ``fatal: not a git repository (or any of the parent directories): .git` I have tried many things ( my gh PAT is correct, my repo is valid, the branch exists ). Can anyone help please?

Hi, what is the eviction policy of task cache (like LRU etc)? Thanks