Open port detection for model deployment

Description

I'm trying to specify a very small range of allowed ports (3 total) for model deployment within steam using the `--prediction-service-port-range=` flag. This seems to be effective, but the model deployment feature from the UI seems to be picking a port at random from the range instead of automatically selecting an open port as is the stated claim.
This means that after I've deployed the first model to a port, the second and third model deployments require several tries and failures before successfully finding an open port.

Assignee

Prithvi Prabhu

Reporter

Kelly O'Briant

Labels

None

AffectedContact

None

AffectedCustomers

None

Priority

Minor
Configure