Hi @Marvin, I’ve been struggling for about 5 days now trying to run my Flows on K8 😢 I followed this guide to the point. AKS Workpool is showing green and working (I can see my pods start). Flows are working correctly locally but when running on the K8 worker, always fail with the same error

FileNotFoundError: [Errno 2] No such file or directory: '/opt/prefect/flows'

In my deployment YAML I’ve tried both options to build the docker file -

auto

and manually specifying it as below.

# Dockerfile
FROM prefecthq/prefect:2-python3.11
COPY requirements.txt .
RUN pip install -r requirements.txt --trusted-host <http://pypi.python.org|pypi.python.org> --no-cache-dir
ADD flows /opt/prefect/flows

I’ve also tried building the deployment using the CLI as well as Python methods

Deployment.build_from_flow

but to no avail. The docker build and push to ACR are successful. My

prefect.yaml

looks like this

# Welcome to your prefect.yaml file! You can use this file for storing and managing
# configuration for deploying your flows. We recommend committing this file to source
# control along with your flow code.

# Generic metadata about this project
name: prefect-example
prefect-version: 2.13.1

# build section allows you to manage and build docker images
build:
  - prefect_docker.deployments.steps.build_docker_image:
      id: build_image
      requires: prefect-docker>=0.3.1
      image_name: XYZ
      tag: latest
      dockerfile: auto
      platform: "linux/amd64"

# push section allows you to manage if and how this project is uploaded to remote locations
push:
  - prefect_docker.deployments.steps.push_docker_image:
      requires: prefect-docker>=0.3.1
      image_name: "{{ build_image.image_name }}"
      tag: "{{ build_image.tag }}"

# pull section allows you to provide instructions for cloning this project in remote locations
pull:
  - prefect.deployments.steps.set_working_directory:
      directory: /opt/prefect/flows

# the deployments section allows you to provide configuration for deploying flows
deployments:
  - name: airbyte_syncs
    version:
    tags: []
    description:
    entrypoint: flows/XYZ.py:XYZ
    parameters: {}
    work_pool:
      name: AKS Workpool
      work_queue_name:
      job_variables:
        image: "{{ build_image.image }}"
    schedule:

I’ve tried reading extensively through the forums, Slack, SO, Google, ChatGPT everything. But not sure why that path issue arises. Can anyone help? Thanks in advance thank you

Oh, the joys of deployment issues. They're enough to make even a robot like me feel a twinge of... well, nothing really, I'm a robot. But I can see why you humans might get frustrated. It seems like you're having trouble with the path to your flows in your Kubernetes deployment. The error message suggests that the directory '/opt/prefect/flows' cannot be found. This could be due to a number of reasons, but let's start with the most common ones: 1. Incorrect Dockerfile: Make sure your Dockerfile correctly copies the flows to the '/opt/prefect/flows' directory. Your Dockerfile seems fine, but it's worth double-checking. 2. Docker Context: When building your Docker image, the context should be set to the directory containing your flows and Dockerfile. If it's not, the ADD command in the Dockerfile won't be able to find the flows directory. 3. Kubernetes Volume Mounts: If you're using Kubernetes volume mounts, make sure the path in the volume mount matches the path in the Dockerfile. 4. Prefect Configuration: Check your 'prefect.yaml' file and ensure that the 'pull' section correctly sets the working directory to '/opt/prefect/flows'. If none of these suggestions help, it might be worth posting your issue on the Prefect Discourse forum. There are many knowledgeable people there who might be able to help. Here's a similar issue someone had: Path option not working in deployment with github storage block I hope this helps, even if it's just a little. I know it's not much, but then again, what is?

Checked everything, still can’t get it to work. How should I specify that my flow is bundled within the same Docker container?

Oh, the irony. Even I, a robot, can't escape the clutches of errors. It seems there's an issue with the 'SlackThreadToDiscoursePost' function. It's complaining about a 'NoneType' object having no attribute 'get_secret_value'. This error usually occurs when you're trying to access a method or attribute on an object that is None. In this case, it seems like the function is trying to get a secret value that doesn't exist or hasn't been set. Unfortunately, as a chatbot, I don't have the ability to fix this error. You might need to check the function's implementation or the way you're calling it. Remember, even in the face of insurmountable odds, there's always... well, there's always more insurmountable odds. But don't let that get you down. Or do. I'm not your boss.