Good day. I am running into some behavior with the prefect CLI that I do not understand. I have a profiles.toml with local, eks, and cloud entries. Even if I specify

prefect profile use 'eks'

or

prefect profile use 'local'

,

prefect deployment ls

returns the list of deployments from my cloud instance. This also occurs with

prefect -p 'eks' deployment ls

or

prefect -p 'local' deployment ls

.

prefect profile ls

shows all three profiles and shows 'eks' as the active profile. The deploy commands also seems stuck on the cloud profile. Any ideas what's going on?

If you run

prefect profile create someprofile

That will create a new profile that can connect to your local sqlite backend. Then do

prefect profile use someprofile

to activate. Then

prefect server start

. Any deployments in the local db should then show up in the UI.

Another possibility could be that you have environment variables that are overriding your profiles (environment variables take precedence over profile settings); you should be able to check this quickly by running

env | grep PREFECT

Thank you both for responding. I have checked the environment variables and there are none related to PREFECT. The server I am trying to (re)connect the Prefect client to is running in a Kubernetes cluster that is accessible right now via forwarding port 4200 to a high number port on my local machine. The contents of the profile therefore look like

active = "eks"

[profiles.eks]
PREFECT_SERVER_API_HOST = "127.0.0.1"
PREFECT_SERVER_API_PORT = "43654"

with

kubectl port-forward svc/prefect-server -n prefect-dev 43654:4200

running in the background. I can connect to the Prefect server dashboard in a web browser through the ingress configured on the cluster (and this method shows a server with access to the API) and also through the forwarded port on localhost (though the localhost relay does not access the API* but I wouldn't expect that since the API should be accessible only to the server except on the forwarded port). Do I need to be forwarding additional ports from my cluster to my local environment for the local client to work? I don't remember needing to do this for setting up previous deployments in the kube deployment but I might be forgetting something, of course.

Pop-up when visiting the remote server dashboard through on the forwarded port, i.e. :
Can't connect to Server API at <https://zzz.yyy.com/api>. Check that it's accessible from your machine.