Hi Metaflow Team, I'm back again. Dumb question: So I'm trying to set up the architecture described here Using this recommended minimal stack as a start. The whole stack is deployed, however, remotely (this required some shoehorning to get to work); and obviously

forward_metaflow_ports.py

won't work as written. Is someone around that has done this before? How do I bind my local metaflow installation to the

metaflow-service

ingress exposed by GKE? This is the first time for me working with any of these tools, so I'm not in top form, sorry. @User back for help :)

good to hear that you got that far 🙂 what's the issue with

forward_metaflow_ports.py

?

nothing is wrong with

forward_metaflow_ports.py

, conceptually. I set up the deployment so that the entire deployment is remote and on GKE, so all ingress to the cluster is inside a google cloud subnet. It seems like I need to put the whole thing behind a load balancer so that users can authenticate and achieve proper port forwarding.

right. When using the port forwarding approach, you can use the normal GCP IAM tools to control access