Hello BentoML Team,

I am unable to deploy my BentoML service to BentoCloud due to a persistent error regarding instance type configuration. Here are the details:

BentoML CLI version: bentoml --version → 1.4.8
Deployment method: Tried both CLI and BentoCloud UI
Service: musicgen_service
Organization: eddiedunn
Cluster: gcp-us-central-1

What I've Tried:

Used CLI: bentoml deploy -n musicgen-service --instance-type cpu.2
Used YAML config (as per docs and UI export): bento: musicgen_service: name: musicgen-service-jgc5 access_authorization: false secrets: [] envs: [] services: apiserver: instance_type: cpu.2 scaling: min_replicas: 0 max_replicas: 1 policy: scale_up_stabilization_window: 0 scale_down_stabilization_window: 600 config_overrides: traffic: timeout: 60 external_queue: false deployment_strategy: Recreate cluster: gcp-us-central-1
Also attempted deployment via the BentoCloud UI wizard.
Error Message (always the same): {"error":"instance type is required for service apiserver"}

Other Info:

My requirements.txt includes boto3 and all other dependencies.
The bento builds and uploads successfully, but the deployment fails at the container startup step.
I see this error in both the CLI and the UI.
Question:
What is the correct, currently supported way to specify the instance type for the apiserver in BentoCloud? Is there a new required field or known backend issue? Please advise on how to proceed.

Thank you!

bentoml deployment apply -f - <<EOF
bento: musicgen_service:ekq4tmq3tgf5nlg6
name: musicgen-service-jgc5
access_authorization: false
secrets: []
envs: []
services:
    apiserver:
        instance_type: cpu.2
        envs: []
        scaling:
            min_replicas: 0
            max_replicas: 1
            policy:
                scale_up_stabilization_window: 0
                scale_down_stabilization_window: 600
        config_overrides:
            traffic:
                timeout: 60
                external_queue: false
        deployment_strategy: Recreate
cluster: gcp-us-central-1

EOF

boto3