I have a question regarding Orion that is confusing me. With Prefect 1.0, we have the concept of flow storage. Is there an equivalent for Orion or how does it work? Say I want to run a flow on k8s, can I just package my flow code in a Docker container along with the required dependencies (maybe using the prefect image as base image)? Thanks for clarifying!

Good question. We are currently working on the storage abstraction. As of now, you can use local storage when you specify

flow_location

on your DeploymentSpec. When you use

DockerFlowRunner

, Prefect will create a Docker container for the flow run and will pull the flow from Orion backend. Similarly when you use

KubernetesFlowRunner

, Prefect will create a Kubernetes job for the flow run and at runtime will pull the flow from the backend. When you create a deployment, you can specify the flag

push_to_server

which is by default set to True. This means that by default, when you create a deployment, the flow text will be loaded from the flow location and stored on the Orion server instead of locally, which allows the flow to be compatible with all flow runners. So in short, by default when you create a deployment, Prefect will store the code in your Orion server, making the process much easier so that you don't need to worry about the storage as much as you had to in Prefect 1.0. Having said that, we are actively working on all that and the best way to stay up-to-date is to watch the #CKNSX5WG3 channel and in the coming weeks you will see more updates on that for sure.

That makes a lot of sense! I will continue watch the channel to look out for updates. Look forward to a production-ready release of Orion so that I can start experimenting with it to replace Argo Workflows 😀

I don't know how this will be handled in Cloud 2.0. I'm pretty sure you will have a lot of flexibility allowing you to choose the right storage method based on your preferences and security/privacy requirements