< Marvin> how to deploy flow without pushing to remote repo Prefect Community #ask-marvin

<@ULVA73B9P> how to deploy flow without pushing to...

Fugo Takefusa

09/18/2025, 1:30 AM

@Marvin how to deploy flow without pushing to remote repo # 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 prefect-version: 3.4.7 # 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: fugotakefusaptml/prefect-docker-repo # tag: 2 # dockerfile: auto # 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/prefect definitions: work_pools: docker_worker_pool: &docker_worker_pool name: docker-worker-pool work_queue_name: null job_variables: image: '{{ build_image.image }}' actions: docker_build: &docker_build - prefect_docker.deployments.steps.build_docker_image: &docker_build_config id: build-image requires: prefect-docker>=0.3.0 image_name: fugotakefusaptml/prefect-docker-repo tag: dev dockerfile: auto docker_push: &docker_push - prefect_docker.deployments.steps.push_docker_image: &docker_push_config requires: prefect-docker>=0.3.1 image_name: fugotakefusaptml/prefect-docker-repo tag: dev # the deployments section allows you to provide configuration for deploying flows deployments: - name: deploy_buy1 flow_name: deploybuy entrypoint: data_flows/deploy_buy.py:buy work_pool: *docker_worker_pool build: *docker_build push: *docker_push

Marvin

09/18/2025, 1:30 AM

✅ thought for 247.5 seconds

Marvin

09/18/2025, 1:35 AM

Great question. In Prefect 3.x you can deploy without pushing to a remote Git repo in a few ways. The right choice depends on your worker setup and whether you’re using Docker. Here are the common patterns: Option A: Local code with a Process work pool (no Docker, no Git) - Create a process pool and start a worker on the same machine where your code lives:

Copy code

prefect work-pool create my-process-pool --type process
prefect worker start --pool my-process-pool

- Deploy from local files (programmatically) using `Flow.from_source`:

Copy code

from prefect import Flow

Flow.from_source(".", "data_flows/deploy_buy.py:buy").deploy(
    name="deploy_buy",
    work_pool_name="my-process-pool",
)

This stores the deployment as “local” storage and expects the worker to read your code from the same path. Option B: Docker work pool with code baked into the image (no Git) - Keep your Docker build step, remove the push step. Your worker must run on the same Docker host so it can see the locally built image. - Edit your prefect.yaml: - Keep

build:

step - Remove

push:

step (or set it to

[]

) - Ensure your work pool

job_variables.image

points at the built image tag - Create/start a Docker worker:

Copy code

uv run --with 'prefect[docker]' prefect work-pool create docker-worker-pool --type docker
uv run --with 'prefect[docker]' prefect worker start --pool docker-worker-pool

- Deploy with

prefect deploy

from your project root, or programmatically. No remote repo is required because the code is inside the image. Notes for your YAML - To skip pushing the image, remove the deployment’s

push: *docker_push

or set

push: []

. - Make sure your Jinja placeholders line up with the step id you set. If your build step id is

build-image

, reference it as

{{ build-image.image }}

job_variables.image

. Option C: Docker work pool with host code mounted into the container (no Git, no image push) - Don’t bake the code into the image; mount it at runtime. - In your Docker work pool job variables, add something like:

Copy code

job_variables:
  image: "python:3.11-slim"
  volumes:
    - "/path/on/host/your-project:/opt/prefect/project"
  working_dir: "/opt/prefect/project"

- In

prefect.yaml

, you can add a pull step to set the working dir (optional if

working_dir

above is set):

Copy code

pull:
  - prefect.deployments.steps.set_working_directory:
      directory: /opt/prefect/project

- Deploy with

prefect deploy

or programmatically. Your worker must be on the same machine (or have the path available) to mount the volume. Quick CLI checks (verified) -

prefect deploy --help

exists and supports entrypoint pointing at

./path/to/file.py:flow_func_name

. -

prefect worker start --pool <name>

and

prefect work-pool create --type <type>

are valid. Docs - Store flow code (Prefect 3.x) Questions to clarify - Are your workers running on the same machine as where you’re deploying? - Do you also want to avoid pushing Docker images to a registry, or just avoid pushing code to a Git repo? - Do you prefer Process workers (no Docker) or Docker workers?

2 Views

Open in Slack

Previous Next