We are setting up flyte in GCP, is ingress and FQDN a mandatory requirement for a production ready setup? can we can run the jobs without having a FQDN?

They are not mandatory, but thr ui uses rest api and cli uses Grpc. If you are using the single binary setup you should be able to Simply port forward

👍

Thanks @freezing-airport-6809 and @freezing-boots-56761 for responding I considered both the advices but still stuck with below problems: 1. @freezing-airport-6809 I see under charts for single binary setup there is eks-starter/production yaml file which looks specific to AWS environment how can I leverage this to deploy it on GCP? 2. @freezing-boots-56761 When I port forward the console service I see below error in UI, which to me looks like it tries to fetch the projects and workflows detail from flyteadmin but it is doing it using the localhost address which is mapped with flyteconsole. Though I am able to do some experiment around task/worlflow registering using cli after portforwarding flyteadmin service

@gifted-train-81198 did you ever find a solution to this? Running into the same problem.

@hallowed-doctor-67759 I was able to utilise the port forwarding inorder to make flyte cli work with remote cluster. For UI console port forwarding didn't help I started using dns

Thanks @gifted-train-81198. I did the same with DNS. Thanks for confirming.

I was able to utilise the port forwarding inorder to make flyte cli work with remote cluster

I have been struggling to get

flytectl

to work with the DNS name and the ingress so I’m going to try this instead since I just got it working on the simple

flyte-binary

deployment.

I also faced error connecting flytectl with dns host, for now I am skipping certificate verification by adding below config in my config.yaml file to make it work

admin:
 # For GRPC endpoints you might want to use dns:///flyte.myexample.com
 insecure: false
 insecureSkipVerify: true