RockteMQ-AI commented on issue #2440:
URL: 
https://github.com/apache/rocketmq-dashboard/issues/2440#issuecomment-5354477776

   **Issue Evaluation**
   
   Category: `bug
   bug` | Status: **Reviewed**
   
   The reported issues are real and reproducible in the current codebase. 
`web/src/pages/home/index.tsx` submits `engine`, `mode`, `model`, and `enhance` 
to `/ai`, but `web/src/pages/ai/chatDraft.ts` and `web/src/pages/ai/index.tsx` 
do not consume `engine` from the draft, so the home-page engine choice is 
dropped. The home page also selects `HOME_MODELS[0]` (`qwen3.8-max`) rather 
than the persisted Studio model on first load. The AI page’s `loadLlmRuntime` 
runs concurrently with draft consumption and can clobber the user-selected 
model. On the backend, `OpenAiCompatibleLlmClient` applies a 60 s total request 
timeout to the streaming request and a second 60 s read timeout, while 
`OpenAiCompatibleLlmGateway` creates a 60 s `SseEmitter`, so slow provider 
startup fails before the first SSE and timeout errors cannot be reliably 
delivered. `AiMessage` renders accumulated chunks directly with 
`ReactMarkdown`, so fragmented SSE events can produce broken code blocks. A 
small stop button is 
 shown during generation but is not prominent.
   The reported issues are real and reproducible in the current codebase. 
`web/src/pages/home/index.tsx` submits `engine`, `mode`, `model`, and `enhance` 
to `/ai`, but `web/src/pages/ai/chatDraft.ts` and `web/src/pages/ai/index.tsx` 
do not consume `engine` from the draft, so the home-page engine choice is 
dropped. The home page also selects `HOME_MODELS[0]` (`qwen3.8-max`) rather 
than the persisted Studio model on first load. The AI page’s `loadLlmRuntime` 
runs concurrently with draft consumption and can clobber the user-selected 
model. On the backend, `OpenAiCompatibleLlmClient` applies a 60 s total request 
timeout to the streaming request and a second 60 s read timeout, while 
`OpenAiCompatibleLlmGateway` creates a 60 s `SseEmitter`, so slow provider 
startup fails before the first SSE and timeout errors cannot be reliably 
delivered. `AiMessage` renders accumulated chunks directly with 
`ReactMarkdown`, so fragmented SSE events can produce broken code blocks. A 
small stop button is 
 shown during generation but is not prominent.
   
   ---
   *Automated evaluation by github-manager*


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to