Repository navigation
10s Timeout #156
|
Hello, I am wondering what the source of the 10s timeout on chat completions is / how I can configure it to be higher. I am just running models locally, so I hit this frequently and I find it doesn't recover so gracefully from this timeout Below is an example, where you can see agent_one generate timing out at 10s, and the children operations finishing only a second later. |
Replies: 2 comments 1 reply
|
That seems unexpectedly low - did that come from running data-analysis with one of the OpenAI models? Can you let me know which so I can see if I can reproduce this. That looks like a hard 10 seconds, and wondering if there's something else (say HTTPX config) that might be causing this. |
|
Just tried this with analysis.py and gpt-4.1 -- not seeing that hard cutoff: |
|
thanks for following up - it seems like this is coming from the mcp inspector I was using, not fast-agent: |
thanks for following up - it seems like this is coming from the mcp inspector I was using, not fast-agent: