Touffman78
Touffman78
Hi @ggozad , My setup is : n8n is on corporate network, with mitmproxy configured on the host to forward HTTPS proxy connexion from n8n to our corporate proxy (HTP-only...
Hi @ggozad , I understand, and hope, your setup is working, as all is local. But my setup is different, and could not be changed : n8n inside corporate network,...
I think this is the same behavior as #15449 but not resolved for Ollama in 1.95.1
Hi @ggozad .... I'm really afraid.... The problem is not Ollama. The problem is the Langchain N8N module for Ollama which is not using the proxy configuration from n8n. As...
Hi, @ggozad After investigating a little in the code and in the doc (my bad : I didn't read this before) I understand that. What is concerning me is :...
> Hey [@Touffman78](https://github.com/Touffman78), after looking at this further, we decided we don't have a way to pass a custom HTTP agent with the proxy to Ollama. I updated the docs...
> I would open an issue with langchain, it's there it needs to happen I think. It occurred to me that another thing to try would be to use the...
It seems axios is not using the 'CONNECT' procedure to the http proxy.
OK, your convinced of.... D But did you test it ? Why the CONNECT method is not used in this case ?
Wow, fast and good work !!!! FI : the issue was raised upgrading from CUDA v12.7 -> v12.8 ! I just installed telegraf_1.34.0~b728c2f7-0 and all is good ! Thx a...