Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

image.png

*How to use custom image for each flow deployment in kubernetes worker pool*

I have prefect k8s work-pool setup with one worker, and I have prefect.yaml setup for deployment. I don't want prefect to build the docker image for me, but I wish to specify each deployment having its own image, is this possible? Currently I configure the
image section of the k8s worker

But this makes all workflow run using this image. Is there a way to adjust my deployment so that different deployment use different images?

```name: flows
prefect-version: 2.14.2

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

# the definitions section allows you to define reusable components for your deployments
definitions:
  tags: &amp;common_tags
    - "sat"
  work_pool: &amp;common_work_pool
    name: "k8s-pool"
    job_variables:
      image: "{{ build-image.image }}"

# the deployments section allows you to provide configuration for deploying flows
deployments:
- name: "flow"
  schedule: null
  entrypoint: "src/sat/sub_flows/flow.py:flow"
  work_pool: *common_work_pool```


I am aware that there is an option with `flow.serve`, I am able to create a deployment for each flow as well, but then I don't know how to roll out new deployment in this way without interrupting a running flow. With flow.serve running in a deployment, when I need to update a flow, I need to delete the deployment and re-deploy that deploymnet, this makes the running flow on original deployment fail. This is way I eneded up switch to use prefect.yaml to specify deployment, and each deployment runs in its own k8s job (I do like this design, much easier and clearer resource separation), but the issue is now I do not know how to set different flow image for each flow in the same k8s worker pool

turns out there is .deploy, so nice! I am guessing through prefect.yaml it can be overwrite by job_variables

```if __name__ == "__main__":
    firstrun_flow.deploy( # type: ignore
        name="example",
        image="custom image here",
        build: False,
        push: False,
    )```

Yep, if you're using `prefect.yaml` and `prefect deploy` you don't want your image built, you can leave the `build` and `push` steps empty and set the `image` for each deployment under `job_variables`.