Hi everyone. I’ve been trying to build a simple graph for a project coordination workflow, and I keep hitting a persistent timeout error when my graph makes an `invoke` call to OpenAI's gpt-4 model.
I’m fairly new to LangGraph and I set up a basic chain with a tool-calling node. The graph runs fine for a few steps, but then it consistently times out on the LLM call. I’m not sure if it’s my configuration, a network issue, or something specific about how LangGraph handles these calls. I’m using the standard `ChatOpenAI` model setup with `gpt-4`.
My setup is pretty simple—just a few nodes and conditional edges. I haven't changed the default timeouts because I’m not sure what a safe value would be. Has anyone else run into this? I’d be grateful for any suggestions on where to start looking or what common fixes might be. I’m a bit overwhelmed by all the possible settings.