and then added these two lines to the end of the flow:

filepaths = get_filepaths(dbt_path, upstream_tasks=[dbt_run])
publish_artifact(filepaths)

Would it make sense to maybe make a PR based on this for the dbt Tasks?

Hi all! Yesterday we released a Task,

FivetranSyncTask

, that lets you kick off a sync with Fivetran and will tell you when that sync has completed (or failed). The goal is to let the user be more deterministic about when to start processing or transforming data that has been landed in a data warehouse by Fivetran. For example, even though Fivetran's UI lets you setup a sync and also setup dbt transformations, we currently don't have the functionality to have the dbt transfromation run after a sync has completed. You could do this now with Prefect and the FivetranSyncTask. This is our very first effort to start bridging the gap between our automated pipelines and the more robust scheduling needs of a Prefect user. It is very early days, we would love feedback! Please reach out to myself and/or @Nick Acosta. We would love to talk to folks who would be interested this, especially to know where this does or does not meet your needs. Thanks!

Please help me how to fix this issue. This is all setup in aws ec2. Prefect is up and running the Python flow every 15 minutes and after 4 hours or so the flow is stopping, UI not accessible. Then I go to ec2 CLI and do "prefect server stop" and I get the below error OpenBLAS blas_thread_init: pthread_create failed for thread 1 of 4: Resource temporarily unavailable OpenBLAS blas_thread_init: RLIMIT_NPROC 63450 current, 63450 max

Hello @Anna Geller (old account) @Jimmy Le I’ve been following your tutorials on running serverless flows using AWS EKS (Fargate), great content and thanks for sharing with us! Anna: https://towardsdatascience.com/distributed-data-pipelines-made-easy-with-aws-eks-and-prefect-106984923b30 Jimmy: https://lejimmy.com/distributed-data-pipelines-with-aws-ecs-fargate-and-prefect-cloud/ I am just a hobbyist Prefect user, and have been using a USD5 / month Ubuntu 20.04 instance to run my flows - scraping data, changing the shape using Pandas, communicating with databases - on a schedule, pretty standard stuff. Deps, config and my own code have been installed onto the instance as a package. Tried out your tutorials out of curiosity and got my flows to work, but couldn’t wrap my head around 2 issues: • Cost - even keeping the Fargate agent running as a serverless EKS cluster, at USD0.10 / hour, would cost about USD72 / month. Is the right way to keep the agent on all the time? • Image size and transfer between runs - I’m using a base image which consists of some pip deps and my own custom package. On flow registration, Prefect creates a new container to copy the flow code in, so if my base image layer is say 500MB, my flow code layer is say 50kb, each serverless run of a flow would still cost 500.050 MB to run as the image is not persistent between flow runs. Each additional flow registered costs the entire size of the base image, plus an additional few kbs of flow code, which seems a bit extravagant While I can appreciate the distributed, fault-tolerant, highly available, self-healing & automatically scalable properties of serverless flow runs, does anyone have any insight on how we can reduce / manage the costs to do so?

Thanks for your responses - I made a viz / summary / cheatsheet following the above guidance for distributed workflows. Hope it is of some help to other people who are interested Prefect + ECS / EKS + S3 / ECR

This message was deleted.

This message was deleted.

This message was deleted.

Ooh, just found that the

toolz

library works pretty seamlessly with Prefect maps. So you can use

pipe

,

compose

,

thread_first

or

thread_last

to chain operations. https://toolz.readthedocs.io/en/latest/api.html#toolz.functoolz.pipe So like, from the Horizontal Mapping blog post, this:

with Flow("mapping-test") as flow:
  sleep.map(sleep.map(times))

cam become this:

with Flow("mapping-test") as flow:
    tz.pipe(times, 
            sleep.map, 
            sleep.map)

which I find a bit more readable! Or lets you turn:

with Flow('iterated map') as flow:
    mapped_result = add_ten.map([1, 2, 3])
    mapped_result_2 = add_ten.map(mapped_result)

into

with Flow('asdasds') as flow:
    mapped_result = tz.pipe([1, 2, 3],
                           add_ten.map,
                           add_ten.map)

Ciao! I'm Jacopo from Coveo! I love prefect and also Metaflow - and I thought to hack together a simple shellTask allowing a Metaflow Flow to live as a task inside prefect (for us, metaflow is typically after dbt and great expectations, so it's a dag inside a dag). I share here the code and 3 mins video of me showing the flow working with the local agent -> Open repo (WIP - don't read too much into it)

<https://github.com/jacopotagliabue/metaflow-as-prefect-task>

Screenshot Overview Here

<https://drive.google.com/file/d/1XoG8UfPpiCSuXYvp9Y27zm6_E4D3C9NX/view>

Open to feedback, especially if you think it's a broad enough use case to deserve a much better polished task class (which can totally be done and I'm happy to work together with others on it!)

Hi everyone! Wanted to share this Prefect flow I wrote to check for COVID-19 vaccine appointments. I found myself checking vaccinespotter.org constantly and that the availability was changing super fast, so I thought “wait I can automate this!” Prefect made it super easy for me to get it on a schedule and setup email notifications. I had it checking very 10 minutes and I was able to find a slot within a day of having it running ⚡

Hi everyone. Some really great content in this channel. I'm new to Prefect and was wondering if there are any open source code bases for real Prefect projects? The above project is nice - but are there larger ones that demonstrate using more Prefect features and best practices?

Hey folks! At Cloud Academy we are true believers of Prefect & dbt so we decided to combine them to build an analytics solution that is now used by our enterprise customers! Here’s the article https://cloudacademy.com/blog/data-engineering-business-intelligence-an-exciting-quest/

Hey people ! I created a Github repository that automates the AWS infrastructure deployment and Agent spin through Github Actions. I’ve also put a GH action in place to register workflows from other repositories. Check my article to learn more about it. https://towardsdatascience.com/introducing-a-dataflow-management-system-backed-up-by-prefect-aws-and-github-actions-3f8c0eef2eb2

Hi community ! My school launched a hackathon this week which aimed to make students experiment big data technologies. After 4 days of development, I'm happy to present you the work my team and I achieved. Working with Prefect to inject COVID-related data and news, it was particularly useful for scheduling daily reports from various sources. We've created a GitHub repo presenting our project "Pandemic Knowledge" : https://github.com/flavienbwk/Pandemic-Knowledge I'd be happy to have feedbacks as well to improve our flows way of coding. Here is a preview of what we got in the end :

@Joe Schmid’s talks are always great and informative — highly recommend!!

I kinda just couldn't resist throwing this meme in here:

Hey Everyone! I don't know how many other people use Prefect for both personal projects and at work. But I put together a small python package that helps make tracking different accounts and switching between them easier. It also has some basic functionality to track different configuration files if you use different settings for different accounts or tokens. Feel free to check it out! https://pypi.org/project/prefect-cloud-manager/0.1.1/

Hello everyone, sharing our event with Heartex about using Prefect to preprocess images before using LabelStudio

I don’t have much to show publicly but just wanna say our rideshare cooperative startup is running a ton of jobs on Prefect and it’s going great 😄 https://www.nytimes.com/2021/05/28/technology/nyc-uber-lyft-the-drivers-cooperative.html

Hey everyone, we have an event with Coiled next week where our where I will demo the new KV Store, and Jeremiah will give general updates about the company.

this GitHub Action will install conda on the runner VM, install Prefect, then register a Flow (using the new

git

Storage to point it to code on a private repo):

# This is a basic workflow to help you get started with Actions

name: CI

# Controls when the action will run. 
on:
  # Triggers the workflow on push or pull request events but only for the main branch
  push:
    branches: [ main ]
  pull_request:
    branches: [ main ]

  # Allows you to run this workflow manually from the Actions tab
  workflow_dispatch:

# A workflow run is made up of one or more jobs that can run sequentially or in parallel
jobs:
  # This workflow contains a single job called "build"
  build:
    # The type of runner that the job will run on
    runs-on: ubuntu-latest

    # Steps represent a sequence of tasks that will be executed as part of the job
    steps:
      - uses: actions/checkout@v2
      - uses: conda-incubator/setup-miniconda@v2
        with:
          auto-update-conda: true
          python-version: 3.8
      - name: install prefect
        shell: bash -l {0}
        run: conda install -c conda-forge prefect -y
      - name: login to Prefect Cloud
        shell: bash -l {0}
        run: prefect auth login -t ${{secrets.PREFECT_TOKEN}}
      - name: Register flow
        shell: bash -l {0}
        run: prefect register --project tester -p test.py -n "run-cloud-fn"

Handy Task that triggers a Google Cloud Function (assumes you have your GCP credentials saved as a

Secret

in Prefect)

import prefect
from prefect import task, Flow, Parameter
from prefect.tasks.secrets.base import PrefectSecret
from google.oauth2 import service_account
from google.auth.transport.requests import AuthorizedSession

@task
def trigger_cloud_fn(
    secret: PrefectSecret, url: str, body: str
):
    credentials = service_account.IDTokenCredentials.from_service_account_info(
        secret, target_audience=url
    )
    authed_session = AuthorizedSession(credentials)
    response = <http://authed_session.post|authed_session.post>(url=url, json=body)
    return response

Cross-posting so this doesn’t get lost

Check out this blog from our partner Slate

Made a new version of the Task that triggers Google Cloud Functions - this one will actually fail if the Cloud Function doesn't work (my last version failed silently)

@task(max_retries=3, retry_delay=datetime.timedelta(seconds=30))
def trigger_cloud_fn(secret: PrefectSecret, url: str, body: t.Dict = dict()):
    logger = prefect.context.get("logger")
    <http://logger.info|logger.info>(body)
    credentials = service_account.IDTokenCredentials.from_service_account_info(
        secret, target_audience=url
    )
    authed_session = AuthorizedSession(credentials)
    response = <http://authed_session.post|authed_session.post>(url=url, json=body)
    if not (isinstance(response, requests.models.Response) and response.ok):
        raise signals.FAIL()
    authed_session.close()
    return response

I was curious if one can put together a 100% open-source equivalent to what is considered Modern Data Stack and wrote a blog to share my findings. Very curious to hear your thoughts and if I missed any great products (of course I didn’t miss P!

Hey all, we recently announced the first release of Open MLOps - an open source production machine learning framework that uses Prefect. • GitHub: https://github.com/datarevenue-berlin/OpenMLOps • Blog announcement: https://datarevenue.com/en-blog/open-mlops-open-source-production-machine-learning TL;DR - it is a set of terraform scripts that sets up Prefect and a bunch of other tools on Kubernetes (currently focusing on EKS / AWS, but should work on others too) as well a set of basic guides and tutorials to get it up and running and build and deploy a basic ML model. Let me know if you try it out 🙂