https://vapi.ai/ logo
Join Discord
Powered by
[WARN] Model request timed out (timeout 767.68ms)
# support
c
We are seeing all our calls from this morning, for each turn are getting '[WARN] Model request timed out' . This is happening after each function call, including 'light' function calls which return very little data/text back to the model' I've tried using gpt-4o and gpt-4-turbo and gpt-3.5-turbo, with and without our own OpenAI key. Is there a way to increase the model timeout?
v
message has been deleted
s
call_id pls
v
@Chris I'm sorry to hear that you're experiencing timeouts with your API calls. Currently, there isn't a direct way to increase the model timeout through the API settings. Timeouts can occur due to various reasons, including network issues, server load, or unusually large requests. Here are a few steps you can take to troubleshoot and potentially mitigate the issue: 1. **Reduce Payload Size**: If possible, try to minimize the size of the data you're sending in each request. Smaller payloads can be processed more quickly, reducing the likelihood of a timeout. 2. **Optimize Requests**: Ensure that your requests are as efficient as possible. This might involve batching requests or pre-processing data to reduce the complexity of the tasks you're asking the model to perform. 3. **Check Network Latency**: Sometimes, the issue might be related to network latency rather than the API itself. Test your network speed and stability to ensure that it's not contributing to the timeouts. 4. **Server Status**: Check OpenAI's status page to see if there are any ongoing issues with the API that might be causing increased response times. 5. **Contact Support**: If none of the above steps resolve the issue, it might be best to reach out to OpenAI support for further assistance. They may be able to provide more insight into the cause of the timeouts and potential solutions. If you'd like, I can assist you in closing this ticket or guide you through contacting support for more detailed help. Would you like to proceed with one of these options?
You can continue this conversation by mentioning me in the message.
c
we have a dev and a live account, from our dev can you check call ID: ae121641-ebd5-4d32-9286-94775a05b010
Wow, thank you for the fast response!
s
In this call_id, I can't see any warning or error message. Can you please confirm with your team whether this is the right call_id or not?
c
let me check again
yes that looks correct, ae121641-ebd5-4d32-9286-94775a05b010
I have other call IDs
s
That's not an issue, you can ignore it.
c
f239b6f6-694d-47b0-8814-603106767b0c
but in the call, it doesn't respond with anything after that
I have to prompt it to respond with the information
s
"fallbackModels": [ "gpt-4-0125-preview", "gpt-4-0613" ] Just remove this part.
c
oh ok
sure
s
Can you tell me exact timestamp in audio?
c
32 seconds
you should be able to hear me re-prompting it to trigger a response
if I don't, I get 'silence-timeout'
it responds with my custom response message "Ok got it"
but then nothing
it's still happening without this
5dbdf6a9-545e-47eb-9b70-1225faa931d5
Hi @Sahil ok looks like I found the issue. It looks like there was a change in the API to the 'messages' object in the function tools which was causing it to not narrate the function call response. This was our function tool object just for your reference: { "name": "confirm_booking", "async": false, "messages": [ { "type": "request-start", "content": "Just give me a minute while I get that for you." }, { "type": "request-complete", "content": "Ok got it. Your booking is confirmed, and you'll receive a confirmation text soon." }, { "type": "request-failed", "content": "I'm having a problem getting that at the moment. Do you want me to try again?" }, { "type": "request-response-delayed", "content": "It appears there is some delay in getting that for you. Please hold on for a moment longer.", "timingMilliseconds": 15000 } ], ... }
we have removed this for now and it's working again
s
Awesome
3 Views
PreviousNext
Powered by