Autocompletion gets super slow

[minwoo-yu]

Until now, I have been using the autocompletion feature through the Codestral API.
However, for the past few days, the autocompletion response has been too slow. (takes about 10 seconds)
I thought this was a problem with Mistral API, but It is same when I tried using local model via ollama remote...

Also, I tried change config file (autocompleteOptions) but it does not helpful...
What's the problem with the autocomplete features?

[Patrick-Erichsen]

Are there any logs you can share?

[minwoo-yu]

Sorry for late, logs are as follows:

[@continuedev] error: The operation was aborted. {"context":"llm_fetch","url":"https://thus-canyon-ozone-dana.trycloudflare.com/api/generate","method":"POST","model":"qwen3:4b","provider":"ollama"}
[@continuedev] error: The operation was aborted. {"context":"llm_complete","model":"qwen3:4b","provider":"ollama","useOpenAIAdapter":false}

Also I tried for Codestral API, the log format was in the same.

[uinstinct]

I tried this with Codestral now. Seems to be working fine with me.

Can you please check it once more on the latest pre-release? Also does it happen on smaller files/codebases as well?

[minwoo-yu]

I think it seems okay with me too.
I'm not sure if there was a minor update, but it's good that it's been fixed.