Vertex AI deploy custom model error - Model server terminated: model server container terminated:

[imozaik]

Hi, 

I'm stuck at following error message when I try to deploy custom model to vertex-ai endpoint.

Command:

gcloud ai endpoints deploy-model {ENDPOINT_ID}\
  --region={REGION} \
  --model={MODEL_ID} \
  --display-name={DEPLOYED_MODEL_NAME} \
  --machine-type=n1-standard-2 \
  --enable-access-logging \
  --enable-container-logging \
  --min-replica-count=1 \
  --max-replica-count=100 \
  --traffic-split=0=100

Error:
Using endpoint [https://europe-west3-aiplatform.googleapis.com/]
ERROR: (gcloud.beta.ai.endpoints.deploy-model) Model server terminated: model server container terminated: exit_code: 0
reason: "Completed"
started_at { seconds: 1668599448 }

finished_at { seconds: 1668599448 } .

The same error appears if I try to deploy from UI.
Note: I was able to import my custom model, and also I can create a new endpoint, but if I try to deploy model or make batch prediction this error appears. Where am I wrong?

Please help me

[anjelab]

You may check here some things you can check further when deploying your model. If this does not work, it would be helpful to file a 1:1 support case since they can check your internal resources.

[WilliamChoy]

Try removing the "--traffic-split" argument, it should be an array.

        --traffic-split=[blue=50, green=50]"

Or add "--deployed-model-id" with a hardcoded value (not "0").