Red Hat build of MicroShift

Verify that your AI model is accessible

Before querying the model through the API, you can verify that the model is accessible and ready to provide answers based on the connected data.

The following examples continue with the OpenVINO Model Server.

Prerequisites

You configured the AI model-serving runtime.
You uploaded your AI model to MicroShift.
MicroShift is running.
You installed OpenShift CLI (oc).

Procedure

Get the IP address of the MicroShift node and assign it to the IP variable as the following example command shows:
```
$ IP=$(oc get nodes -o json | jq -r '.items[0].status.addresses[0].address')
```
Identify the name of the route you created by running the following command:
```
$ oc get route -n ai-test <route_name> -o yaml
```
where:

<route_name>

Specifies the actual name of your route.
Extract and assign the HOST value of the route to the DOMAIN variable by running the following command:
```
DOMAIN=$(oc get route -n ai-test <route_name> -o=jsonpath="{ .status.ingress[0].host }")
```
where:

<route_name>

Specifies the actual name of your route.

Enable data transfer from the route to the MicroShift IP address by running the following command:

$ curl -i "${DOMAIN}/v2/models/ovms-resnet50/ready" --connect-to "${DOMAIN}::${IP}:"

Instead of using the --connect-to "${DOMAIN}::${IP}:" flag, you can also use real DNS, or add the IP address and the domain to the /etc/hosts file.

Example output

HTTP/1.1 200 OK
content-type: application/json
date: Wed, 12 Mar 2025 16:01:32 GMT
content-length: 0
set-cookie: 56bb4b6df4f80f0b59f56aa0a5a91c1a=4af1408b4a1c40925456f73033d4a7d1; path=/; HttpOnly

Query the model metadata by running the following command:

$ curl "${DOMAIN}/v2/models/ovms-resnet50" --connect-to "${DOMAIN}::${IP}:"

Example output

{"name":"ovms-resnet50","versions":["1"],"platform":"OpenVINO","inputs":[{"name":"0","datatype":"FP32","shape":[1,224,224,3]}],"outputs":[{"name":"1463","datatype":"FP32","shape":[1,1000]}]

Next steps

Verify that your model is ready for inferencing.
Query the model.
Verify the model response.
Optional: Get the model server metrics.