bash
   prefect work-pool create <work_pool_name> --type <pool_type>

python
   from prefect_azure import AzureContainerInstanceJob
   from prefect_azure_credentials import AzureContainerInstanceCredentials

   container_instance_job = AzureContainerInstanceJob(
       aci_credentials=AzureContainerInstanceCredentials.load("MY_BLOCK_NAME"),
       resource_group_name="azure_resource_group.example.name",
       subscription_id="<MY_AZURE_SUBSCRIPTION_ID>",
   )
   container_instance_job.save("aci-dev")

bash
   prefect deployment build a_flow_module.py:log_hello_flow --name aci-dev -ib container-instance-job/aci-dev

bash
   prefect work-pool create -t azure-container-instance my-aci-work-pool

bash
   prefect worker start -n my-aci-worker -p my-aci-work-pool

python
from prefect_azure.container_instance import ACI_DEFAULT_CPU, ACI_DEFAULT_MEMORY

container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("MY_BLOCK_NAME"),
    resource_group_name="azure_resource_group.example.name",
    subscription_id="<MY_AZURE_SUBSCRIPTION_ID>",
    cpu=2.0,  # Number of virtual CPUs
    memory=4.0  # Amount of memory in GB
)
container_instance_job.save("aci-dev")

cpu

memory

bash
   prefect work-pool create <work_pool_name> --type push

cpu

memory

python
   from prefect_azure import AzureContainerInstanceJob
   from prefect_azure_credentials import AzureContainerInstanceCredentials

   container_instance_job = AzureContainerInstanceJob(
       aci_credentials = AzureContainerInstanceCredentials.load("MY_BLOCK_NAME"),
       resource_group_name = "azure_resource_group.example.name",
       subscription_id = "<MY_AZURE_SUBSCRIPTION_ID>",
       cpu = 2.0,  # Number of virtual CPUs
       memory = 4.0  # Amount of memory in GB
   )
   container_instance_job.save("aci-dev")

bash
   prefect deployment build flow.py:my_flow --name aci-dev -ib container-instance-job/aci-dev

bash
   prefect worker start -n my-aci-worker -p <work_pool_name>

cpu

memory

AzureContainerInstanceJob

python
from prefect_azure import AzureContainerInstanceJob
from prefect_azure_credentials import AzureContainerInstanceCredentials

# Load your existing credentials block
aci_credentials = AzureContainerInstanceCredentials.load("MY_BLOCK_NAME")

# Create the job with additional resource specifications
container_instance_job = AzureContainerInstanceJob(
    aci_credentials=aci_credentials,
    resource_group_name="azure_resource_group.example.name",
    subscription_id="<MY_AZURE_SUBSCRIPTION_ID>",
    cpu=2.0,  # Number of virtual CPUs you want to allocate
    memory=4.0  # Amount of memory in GB you want to allocate
)

# Save the updated job configuration
container_instance_job.save("aci-dev")

bash
prefect deployment build flow.py:my_flow --name aci-dev -ib container-instance-job/aci-dev

bash
prefect worker start -n my-aci-worker -p <work_pool_name>

alright, but got another error on s3fs since I’m using s3 bucket for storage. even though I’m sure I’m installing the s3fs package within the docker container. but still getting this error: Flow could not be retrieved from deployment.

alright, but got another error on s3fs since I’m using s3 bucket for storage. even though I’m sure I’m installing the s3fs package within the docker container. but still getting this error: Flow could not be retrieved from deployment.

alright, but got another error on s3fs since I’m using s3 bucket for storage. even though I’m sure I’m installing the s3fs package within the docker container. but still getting this error: Flow could not be retrieved from deployment.

ModuleNotFoundError: No module named 's3fs'

s3fs

s3fs

s3fs

s3fs

s3fs

requirements.txt

setup.py

txt
   # requirements.txt
   s3fs==2021.11.1  # replace with your required version

s3fs

Dockerfile
   FROM prefecthq/prefect:latest-python3.9  # Example base image

   # Install s3fs
   RUN pip install s3fs

s3fs

bash
   docker build -t my-prefect-image .

s3fs

bash
   docker run -it my-prefect-image
   python -c "import s3fs; print(s3fs.__version__)"

pip

pip check

bash
  pip check

bash
  pip install s3fs==2021.11.1 botocore==1.27.19 boto3==1.24.19

S3

s3fs

s3fs

s3fs

s3fs

cmd
docker exec -it <container_id> cmd
pip show s3fs

s3fs

s3fs

cmd
docker exec -it <container_id> cmd
pip install --force-reinstall s3fs

Dockerfile
FROM prefecthq/prefect:latest

# Install s3fs
RUN pip install s3fs

bash
docker build -t my-prefect-linux-image .
docker run -it my-prefect-linux-image

s3fs

bash
   pip install -U prefect-azure

bash
   prefect block register -m prefect_azure

python
from prefect_azure import AzureBlobStorage
from prefect.blocks.system import JSON

# Create a block with your Azure Blob Storage credentials
blob_storage_block = AzureBlobStorage(
    container="my-container",
    connection_string="DefaultEndpointsProtocol=https;AccountName=myaccountname;AccountKey=myaccountkey;EndpointSuffix=<http://core.windows.net|core.windows.net>"
)

# Save the block
blob_storage_block.save("my-azure-blob-storage")

python
from prefect.deployments import Deployment
from my_flow import my_flow_function

deployment = Deployment.build_from_flow(
    flow=my_flow_function,
    name="my-flow-deployment",
    storage=AzureBlobStorage.load("my-azure-blob-storage"),
)

deployment.apply()

asyncio - Unclosed client session

asyncio - Unclosed client session

async with

async with

python
import os
import asyncio
from prefect import flow
from prefect_azure import AzureBlobStorageCredentials

@flow
async def example_get_client_flow():
    connection_string = os.getenv("AZURE_STORAGE_CONNECTION_STRING")
    azure_credentials = AzureBlobStorageCredentials(
        connection_string=connection_string,
    )
    async with azure_credentials.get_client() as blob_service_client:
        # run other code here
        pass

asyncio.run(example_get_client_flow())

python
from prefect import flow
from prefect_azure import AzureBlobStorage

@flow
async def my_flow():
    async with AzureBlobStorage.load("my-azure-blob-storage") as blob_storage:
        # Your code to interact with Azure Blob Storage
        pass

    # Ensure all sessions are closed
    await blob_storage.close()

asyncio.run(my_flow())

asyncio - Unclosed client session

getting this error running my flow: Flow could not be retrieved from deployment.

Flow could not be retrieved from deployment.

Flow could not be retrieved from deployment.

KeyError: “No class found for dispatch key ‘azure-blob-storage-container’ in registry for type ‘Block’.”

azure-blob-storage-container

prefect-azure

prefect-azure

bash
   pip install -U prefect-azure
   prefect block register -m prefect_azure

azure-blob-storage-container

Dockerfile
   FROM prefecthq/prefect:latest

   # Install prefect-azure
   RUN pip install -U prefect-azure

   # Register the blocks
   RUN prefect block register -m prefect_azure

bash
   docker build -t my-prefect-image .
   docker run -it my-prefect-image prefect block register -m prefect_azure

azure-blob-storage-container

python
   from prefect_azure import AzureBlobStorage

   # Example of creating and using Azure Blob Storage Block
   blob_storage_block = AzureBlobStorage(
       container="my-container",
       connection_string="DefaultEndpointsProtocol=https;AccountName=myaccountname;AccountKey=myaccountkey;EndpointSuffix=<http://core.windows.net|core.windows.net>"
   )

   # Save the block
   blob_storage_block.save("my-azure-blob-storage")

python
   from prefect import flow
   from prefect_azure import AzureBlobStorage

   @flow
   def my_flow():
       blob_storage = AzureBlobStorage.load("my-azure-blob-storage")
       # Add your flow logic here

   my_flow()

python
import asyncio
from prefect import flow
from prefect_azure import AzureBlobStorageCredentials

@flow
async def example_get_client_flow():
    connection_string = os.getenv("AZURE_STORAGE_CONNECTION_STRING")
    azure_credentials = AzureBlobStorageCredentials(
        connection_string=connection_string,
    )
    async with azure_credentials.get_client() as blob_service_client:
        # run other code here
        pass

asyncio.run(example_get_client_flow())

KeyError

azure-blob-storage-container

prefect-gams-windows:latest

prefect-gams-windows:latest

bash
   az acr create --resource-group myResourceGroup --name myRegistry --sku Basic

bash
   az acr login --name myRegistry

bash
   docker tag prefect-gams-windows:latest <http://myRegistry.azurecr.io/prefect-gams-windows:latest|myRegistry.azurecr.io/prefect-gams-windows:latest>
   docker push <http://myRegistry.azurecr.io/prefect-gams-windows:latest|myRegistry.azurecr.io/prefect-gams-windows:latest>

python
   from prefect_azure import AzureContainerRegistryCredentials

   acr_credentials = AzureContainerRegistryCredentials(
       registry_url="<http://myRegistry.azurecr.io|myRegistry.azurecr.io>",
       username="myRegistryUsername",
       password="myRegistryPassword"
   )
   acr_credentials.save("my-azure-container-registry")

python
   from prefect.blocks.system import DockerContainer

   docker_container_block = DockerContainer(
       image="<http://myRegistry.azurecr.io/prefect-gams-windows:latest|myRegistry.azurecr.io/prefect-gams-windows:latest>",
       registry_credentials=AzureContainerRegistryCredentials.load("my-azure-container-registry")
   )
   docker_container_block.save("my-docker-container")

my-docker-container

python
from prefect.blocks.system import DockerRegistry

docker_registry_block = DockerRegistry(
    registry_url="<http://index.docker.io|index.docker.io>",
    username="myDockerHubUsername",
    password="myDockerHubPassword"
)
docker_registry_block.save("my-docker-hub-registry")

docker_container_block = DockerContainer(
    image="myDockerHubUsername/prefect-gams-windows:latest",
    registry_credentials=DockerRegistry.load("my-docker-hub-registry")
)
docker_container_block.save("my-docker-container")

<http://myRegistry.azurecr.io|myRegistry.azurecr.io>

AzureContainerRegistryCredentials

bash
   az ad sp create-for-rbac --name acr-pull --role Contributor --scopes /subscriptions/{subscription_id}/resourceGroups/{resource_group}/providers/Microsoft.ContainerRegistry/registries/{acr_name}

appId

password

tenant

python
   from prefect.blocks.system import DockerRegistry

   acr_credentials = DockerRegistry(
       registry_url="<http://myRegistry.azurecr.io|myRegistry.azurecr.io>",
       username="clientId",
       password="clientSecret"
   )
   acr_credentials.save("my-acr-registry-credentials")

python
   from prefect.blocks.system import DockerContainer

   docker_container_block = DockerContainer(
       image="<http://myRegistry.azurecr.io/prefect-gams-windows:latest|myRegistry.azurecr.io/prefect-gams-windows:latest>",
       registry_credentials=DockerRegistry.load("my-acr-registry-credentials")
   )
   docker_container_block.save("my-docker-container")

my-docker-container

python
from prefect.blocks.system import DockerRegistry, DockerContainer

# Define and save the Docker registry credentials block
acr_credentials = DockerRegistry(
    registry_url="<http://myRegistry.azurecr.io|myRegistry.azurecr.io>",
    username="clientId",        # Output from `az ad sp create-for-rbac`
    password="clientSecret"     # Output from `az ad sp create-for-rbac`
)
acr_credentials.save("my-acr-registry-credentials")

# Define and save the Docker container block
docker_container_block = DockerContainer(
    image="<http://myRegistry.azurecr.io/prefect-gams-windows:latest|myRegistry.azurecr.io/prefect-gams-windows:latest>",
    registry_credentials=DockerRegistry.load("my-acr-registry-credentials")
)
docker_container_block.save("my-docker-container")

python
from prefect.deployments import Deployment

deployment = Deployment.build_from_flow(
    flow=my_flow_function,  # Replace with your flow function
    name="my-flow-deployment",
    infrastructure=DockerContainer.load("my-docker-container")
)

deployment.apply()

Windows

yaml
apiVersion: 2018-10-01
location: eastus
name: myACI
properties:
  osType: Windows  # Change this to Windows
  containers:
    - name: mycontainer
      image: <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>
      resources:
        requests:
          cpu: 1.0
          memoryInGb: 1.5

python
from prefect_azure.container_instance import AzureContainerInstanceJob, AzureContainerInstanceCredentials

azure_credentials_block = AzureContainerInstanceCredentials(
    resource_group_name="myResourceGroup",
    subscription_id="<my_subscription_id>",
    tenant_id="<my_tenant_id>",
    client_id="<my_client_id>",
    client_secret="<my_client_secret>"
)
azure_credentials_block.save("my-azure-credentials")

azure_container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("my-azure-credentials"),
    region="eastus",
    osType="Windows",  # Specify the OS as Windows here
    container_resource_requests=dict(cpu=1.0, memory=1.5),
    container_name="mycontainer",
    image="<http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>"
)

azure_container_instance_job.save("my-azure-job")

osType

osType

additional_config

osType

osType

python
from prefect_azure.container_instance import AzureContainerInstanceJob, AzureContainerInstanceCredentials

# Create credentials block
azure_credentials_block = AzureContainerInstanceCredentials(
    resource_group_name="myResourceGroup",
    subscription_id="my_subscription_id",
    tenant_id="my_tenant_id",
    client_id="my_client_id",
    client_secret="my_client_secret"
)
azure_credentials_block.save("my-azure-credentials")

# Create Azure Container Instance Job with Windows OS Type
azure_container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("my-azure-credentials"),
    region="eastus",
    additional_config={
        "osType": "Windows"
    },
    container_resource_requests={"cpu": 1.0, "memory": 1.5},
    container_name="mycontainer",
    image="<http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>"
)
azure_container_instance_job.save("my-azure-aci-job")

python
from prefect.deployments import Deployment
from my_flow import my_flow_function  # Replace with your actual flow import

# Assuming your flow is defined in my_flow_function
deployment = Deployment.build_from_flow(
    flow=my_flow_function,
    name="my-flow-deployment",
    infrastructure=AzureContainerInstanceJob.load("my-azure-aci-job")
)

deployment.apply()

AzureContainerInstanceJob

osType

Windows

osType

Windows

additional_config

AzureContainerInstanceJob

region

AzureContainerInstanceJob

AzureContainerInstanceJob

python
from prefect_azure.container_instance import AzureContainerInstanceJob, AzureContainerInstanceCredentials

# Create or load Azure credentials block
azure_credentials_block = AzureContainerInstanceCredentials(
    resource_group_name="myResourceGroup",
    subscription_id="my_subscription_id",
    tenant_id="my_tenant_id",
    client_id="my_client_id",
    client_secret="my_client_secret"
)
azure_credentials_block.save("my-azure-credentials")

# Create Azure Container Instance Job with the specified region
azure_container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("my-azure-credentials"),
    region="eastus",  # Specify the Azure region here
    additional_config={"osType": "Windows"},  # Ensure osType is set to Windows if needed
    container_resource_requests={"cpu": 1.0, "memory": 1.5},
    container_name="mycontainer",
    image="<http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>"
)
azure_container_instance_job.save("my-azure-aci-job")

python
from prefect.deployments import Deployment
from my_flow import my_flow_function  # Replace with your actual flow import

deployment = Deployment.build_from_flow(
    flow=my_flow_function,
    name="my-flow-deployment",
    infrastructure=AzureContainerInstanceJob.load("my-azure-aci-job")
)

deployment.apply()

region

AzureContainerInstanceJob

guest RPC failure: The system cannot find the file specified.: unknown

C:\\path\\to\\file

bash
   docker build -t <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest> .
   docker push <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>

Dockerfile
# Use a base Windows image
FROM <http://mcr.microsoft.com/windows:latest|mcr.microsoft.com/windows:latest>

# Set environment variables if needed
ENV VAR_NAME value

# Copy necessary files
COPY . /app

# Install any dependencies
RUN powershell -Command Set-ExecutionPolicy Bypass -Scope Process -Force; `
    [Net.ServicePointManager]::SecurityProtocol = [Net.SecurityProtocolType]::Tls12; `
    Invoke-WebRequest -Uri <https://path/to/dependency> -OutFile C:\path\to\dependency

# Set the working directory
WORKDIR /app

# Specify the entrypoint or command
ENTRYPOINT ["powershell", "-Command", "path\\to\\entrypoint.ps1"]

python
from prefect_azure.container_instance import AzureContainerInstanceJob, AzureContainerInstanceCredentials

# Create or load Azure credentials block
azure_credentials_block = AzureContainerInstanceCredentials(
    resource_group_name="myResourceGroup",
    subscription_id="my_subscription_id",
    tenant_id="my_tenant_id",
    client_id="my_client_id",
    client_secret="my_client_secret"
)
azure_credentials_block.save("my-azure-credentials")

# Create Azure Container Instance Job with the specified region and Windows OS
azure_container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("my-azure-credentials"),
    region="eastus",  # Specify the Azure region
    additional_config={"osType": "Windows"},
    container_resource_requests={"cpu": 1.0, "memory": 1.5},
    container_name="mycontainer",
    image="<http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>"
)
azure_container_instance_job.save("my-azure-aci-job")

"/opt/prefect/entrypoint.sh"

/opt/prefect/entrypoint.sh

entrypoint.ps1

powershell
# entrypoint.ps1

# Activate environment if using a virtualenv or conda, e.g.:
# & "C:\path\to\venv\Scripts\Activate.ps1"

# Ensure Prefect is installed and available
pip install prefect

# Run the Prefect agent or worker
prefect agent start -q "my-work-queue"

Dockerfile
# Use a base Windows image
FROM <http://mcr.microsoft.com/windows/servercore:ltsc2019|mcr.microsoft.com/windows/servercore:ltsc2019>

# Set environment variables if needed
ENV VAR_NAME value

# Install Python and Prefect
# Ensure that this matches the correct path and Python version
RUN powershell -Command `
    Invoke-WebRequest -Uri <https://www.python.org/ftp/python/3.9.7/python-3.9.7-amd64.exe> -OutFile python-3.9.7-amd64.exe; `
    Start-Process python-3.9.7-amd64.exe -ArgumentList '/quiet InstallAllUsers=1 PrependPath=1' -NoNewWindow -Wait; `
    Remove-Item -Force python-3.9.7-amd64.exe; `
    pip install --upgrade pip; `
    pip install prefect

# Copy your flow and other necessary files
COPY . /app

# Set the working directory
WORKDIR /app

# Copy the entrypoint script
COPY entrypoint.ps1 .

# Set the entrypoint to the PowerShell script
ENTRYPOINT ["powershell.exe", "C:\\app\\entrypoint.ps1"]

bash
docker build -t <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest> .
docker push <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>

python
from prefect_azure.container_instance import AzureContainerInstanceJob, AzureContainerInstanceCredentials

# Create or load Azure credentials block
azure_credentials_block = AzureContainerInstanceCredentials(
    resource_group_name="myResourceGroup",
    subscription_id="my_subscription_id",
    tenant_id="my_tenant_id",
    client_id="my_client_id",
    client_secret="my_client_secret"
)
azure_credentials_block.save("my-azure-credentials")

# Create Azure Container Instance Job with the specified region and Windows OS
azure_container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("my-azure-credentials"),
    region="eastus",  # Specify the Azure region
    additional_config={"osType": "Windows"},
    container_resource_requests={"cpu": 1.0, "memory": 1.5},
    container_name="mycontainer",
    image="<http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>"
)
azure_container_instance_job.save("my-azure-aci-job")

prefect worker

prefect agent

entrypoint.ps1

powershell
# entrypoint.ps1

# Activate environment if using a virtualenv or conda, e.g.:
# & "C:\path\to\venv\Scripts\Activate.ps1"

# Ensure Prefect is installed and available
pip install prefect

# Start the Prefect worker for the specified work pool
prefect worker start -n "my-worker-name" -p "my-push-work-pool"

Dockerfile
# Use a base Windows image
FROM <http://mcr.microsoft.com/windows/servercore:ltsc2019|mcr.microsoft.com/windows/servercore:ltsc2019>

# Set environment variables if needed
ENV VAR_NAME value

# Install Python and Prefect
# Ensure this matches the correct path and Python version
RUN powershell -Command `
    Invoke-WebRequest -Uri <https://www.python.org/ftp/python/3.9.7/python-3.9.7-amd64.exe> -OutFile python-3.9.7-amd64.exe; `
    Start-Process python-3.9.7-amd64.exe -ArgumentList '/quiet InstallAllUsers=1 PrependPath=1' -NoNewWindow -Wait; `
    Remove-Item -Force python-3.9.7-amd64.exe; `
    pip install --upgrade pip; `
    pip install prefect

# Copy your flow and other necessary files
COPY . /app

# Set the working directory
WORKDIR /app

# Copy the entrypoint script
COPY entrypoint.ps1 .

# Set the entrypoint to the PowerShell script
ENTRYPOINT ["powershell.exe", "C:\\app\\entrypoint.ps1"]

bash
docker build -t <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest> .
docker push <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>

AzureContainerInstanceJob

python
from prefect_azure.container_instance import AzureContainerInstanceJob, AzureContainerInstanceCredentials

# Create or load Azure credentials block
azure_credentials_block = AzureContainerInstanceCredentials(
    resource_group_name="myResourceGroup",
    subscription_id="my_subscription_id",
    tenant_id="my_tenant_id",
    client_id="my_client_id",
    client_secret="my_client_secret"
)
azure_credentials_block.save("my-azure-credentials")

# Create Azure Container Instance Job with the specified region and Windows OS
azure_container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("my-azure-credentials"),
    region="eastus",  # Specify the Azure region
    additional_config={"osType": "Windows"},
    container_resource_requests={"cpu": 1.0, "memory": 1.5},
    container_name="mycontainer",
    image="<http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>"
)
azure_container_instance_job.save("my-azure-aci-job")

python
from prefect.deployments import Deployment
from my_flow import my_flow_function  # Replace with your actual flow import

deployment = Deployment.build_from_flow(
    flow=my_flow_function,
    name="my-flow-deployment",
    infrastructure=AzureContainerInstanceJob.load("my-azure-aci-job")
)

deployment.apply()

entrypoint.ps1

powershell
# entrypoint.ps1

# Set up the environment if needed
# Activate environment if using a virtualenv or conda, e.g.:
# & "C:\path\to\venv\Scripts\Activate.ps1"

# Ensure Prefect is installed and available
pip install prefect

# This script might include steps to initialize or start specific commands
# since push work pools handle execution without explicit worker commands.

Dockerfile
# Use a base Windows image
FROM <http://mcr.microsoft.com/windows/servercore:ltsc2019|mcr.microsoft.com/windows/servercore:ltsc2019>

# Set environment variables if needed
ENV VAR_NAME value

# Install Python and Prefect
# Ensure this matches the correct path and Python version
RUN powershell -Command `
    Invoke-WebRequest -Uri <https://www.python.org/ftp/python/3.9.7/python-3.9.7-amd64.exe> -OutFile python-3.9.7-amd64.exe; `
    Start-Process python-3.9.7-amd64.exe -ArgumentList '/quiet InstallAllUsers=1 PrependPath=1' -NoNewWindow -Wait; `
    Remove-Item -Force python-3.9.7-amd64.exe; `
    pip install --upgrade pip; `
    pip install prefect

# Copy your flow and other necessary files
COPY . /app

# Set the working directory
WORKDIR /app

# Copy the entrypoint script
COPY entrypoint.ps1 .

# Set the entrypoint to the PowerShell script
ENTRYPOINT ["powershell.exe", "C:\\app\\entrypoint.ps1"]

bash
docker build -t <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest> .
docker push <http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>

python
from prefect_azure.container_instance import AzureContainerInstanceJob, AzureContainerInstanceCredentials

# Create or load Azure credentials block
azure_credentials_block = AzureContainerInstanceCredentials(
    resource_group_name="myResourceGroup",
    subscription_id="my_subscription_id",
    tenant_id="my_tenant_id",
    client_id="my_client_id",
    client_secret="my_client_secret"
)
azure_credentials_block.save("my-azure-credentials")

# Create Azure Container Instance Job with the specified region and Windows OS
azure_container_instance_job = AzureContainerInstanceJob(
    aci_credentials=AzureContainerInstanceCredentials.load("my-azure-credentials"),
    region="eastus",  # Specify the Azure region
    additional_config={"osType": "Windows"},
    container_resource_requests={"cpu": 1.0, "memory": 1.5},
    container_name="mycontainer",
    image="<http://my-acr.azurecr.io/prefect-gams-windows:latest|my-acr.azurecr.io/prefect-gams-windows:latest>"
)
azure_container_instance_job.save("my-azure-aci-job")

python
from prefect.deployments import Deployment
from my_flow import my_flow_function  # Replace with your actual flow import

deployment = Deployment.build_from_flow(
    flow=my_flow_function,
    name="my-flow-deployment",
    infrastructure=AzureContainerInstanceJob.load("my-azure-aci-job")
)

deployment.apply()