Inference API returns an incorrect error message when inference ID = model ID #111312
Description
Elasticsearch Version
8.15
Installed Plugins
No response
Java Version
bundled
OS Version
Mac OS
Problem Description
When creating an inference endpoint using an inference_id that is identical to the model_id, it returns a misleading error suggesting that the inference API is trying to re-deploy the model again after eland has already deployed it to ML.
Steps to Reproduce
When creating an inference endpoint using an inference_id that is identical to the model_id:
PUT _inference/text_embedding/sentence-transformers__all-minilm-l6-v2
{
"service": "elasticsearch",
"service_settings": {
"num_allocations": 1,
"num_threads": 1,
"model_id": "sentence-transformers__all-minilm-l6-v2"
}
}
It returns a misleading error message:
{
"error": {
"root_cause": [
{
"type": "status_exception",
"reason": "Model IDs must be unique. Requested model ID [sentence-transformers__all-minilm-l6-v2] matches existing model IDs but must not."
}
],
"type": "status_exception",
"reason": "Model IDs must be unique. Requested model ID [sentence-transformers__all-minilm-l6-v2] matches existing model IDs but must not."
},
"status": 400
}
The error should say something like the following instead:
Inference ID [<inference_id>] must be unique and must not match the <model_id>.
Logs (if relevant)
No response