custom name for openai api connection breaks pipeline

[mamema]

Self Checks

• [x] This is only for bug report, if you would like to ask a question, please head to Discussions.
• [x] I have searched for existing issues search for existing issues, including closed ones.
• [x] I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [x] [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• [x] Please do not modify this template :) and fill in all the required fields.

Dify version

1.0.1

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

i have two diffrent GPU modell provider endpoints which i want to utlizie in dify for creation e of an openai api end points three main config parameters are needed: 1.) Model name 2.) API endpoint url 3.) model name for API endpoint

i am not able to enter twice the same for 1.) and 2.) but with a different endpoint ip address for 3.) it would be difficult during selection process, as the names are both the same. Only one unique name per model is allowed. It is not possible to utilize two endpoints with the same model

So i´ve entered different names. like shown in the screenshots

it accepts the modified names, as shown in the screenshots, and it´s even worse, it is showing working ingestion in knowledge stacks. But as soon as within a ChatBot the knowledge stack i connected and retrieval parameters are set (in my case Hybrid search- Weighted) and a question is asked to the llm model (llama 3.1:8b-instruct), this error is shown:

The error is NOT shown, when i chat with the model without attaching a knowledge stack

✔️ Expected Behavior

custom names should be possible to connect multiple provider endpoints within one plugin (openai compatible providers in my example) without breaking the RAG pipeline

❌ Actual Behavior

RAG pipeline breaks and chatting with documents isn´t possible