Using "text-embedding-3-small" - please provide the correct setting with V4

Support
1

Description

I moved to “text-embedding-3-small” and there is probably issue with insert, what to exactly, way to use, I use azure

vectorizer_config_ = [
Configure.NamedVectors.text2vec_azure_openai(
base_url=settings.OPENAI_AZURE_MODEL_TEXT_3_SMALL_ENDPOINT,
deployment_id=settings.OPENAI_AZURE_MODEL_TEXT_3_SMALL_MODEL,
model=settings.OPENAI_AZURE_MODEL_TEXT_3_SMALL_MODEL,
resource_name=settings.OPENAI_AZURE_MODEL_TEXT_3_SMALL_VERSION,
name=“_”.join(item) if isinstance(item, list) else item,
s

2

model=settings.OPENAI_AZURE_MODEL_TEXT_3_SMALL_MODEL,
just added and tried - but it seems not exists in the version I have, not sure this is the problem

3

Good morning @dror.pipano,

Welcome to our community—it’s lovely to have you here! :hugs:

Here’s how to integrate Weaviate with Azure OpenAI for embeddings:

I am not expect in Azure deployment however:

  1. In Azure, I think you need first deploy the embedding model you want to use (e.g., text-embedding-3-small) in Azure side of things.
  2. Once deployed, configure Weaviate with the following parameters:
  • resource_name: The name of your Azure OpenAI resource
  • deployment_id: The deployment ID of your model in Azure
  • base_url: Your Azure endpoint URL (e.g., https://.openai.azure.com)

Here are some helpful Azure documentation for reference:

Best regards,

Mohamed Shahin
Weaviate Support Engineer
(Ireland, GMT/UTC timezone)

4

I have the resource works fine on simple
notebook,. - but get this error: ERROR | 2025-05-06 01:50:54 | weaviate-client.__send_batch | None | {‘message’: ‘Failed to send all objects in a batch of 1’, ‘error’: “WeaviateBatchError(‘Query call with protocol GRPC batch failed with message Deadline Exceeded.’)”}

5

The “Deadline Exceeded” error typically indicates a timeout, meaning the server was unable to fulfill the request within the time window from the client.

I recommend the following:

  1. Increase the timeout values in your Weaviate connection method to allow more time.
  2. Reduce the batch size, and if you’re currently using a dynamic size, switch to a fixed, smaller batch for more consistent performance.
  3. Check your resource usage—ensure there are no CPU, memory, or network bottlenecks on the server side that could be causing delays.

Best regards,
Mohamed Shahin
Weaviate Support Engineer
(Ireland, GMT/UTC timezone)

6

I think this would be the root cause, this is a test for a single line, perhaps it’s a wait time for azure resource, but this works fine with jupyter simple test..

1 Like