Everytime i try to switch to gpt 4o cluster vapi c...
# support
r
v
Action Step: Since the error persists when switching to the gpt-4o cluster, please contact Vapi support for direct assistance. You can email support@vapi.ai or join the [Vapi Discord community](https://discord.com/invite/pUFNcf2WmH) for real-time help. Tip: Include the error message and a screenshot when reaching out to speed up troubleshooting. Source: - [Vapi Support](https://docs.vapi.ai/support)
c
Hi Ronan, If you're experiencing crashes when switching to the gpt-4o model, ensure you're specifying the model correctly in your configuration. According to the [October 8, 2024 changelog](https://docs.vapi.ai/changelog/2024/10/8), you should use
gpt-4o-2024-08-06
for Azure OpenAI. this appears to be related to concurrency and capacity management:** 1. Common Causes: - Exceeding organization's concurrent call limit - Too many simultaneous WebSocket connections - Backend resource constraints 2. How to Address: - Check your organization's concurrency limits - Space out call attempts (avoid sending too many at once) - Monitor active call count - Implement backoff strategies for retries 3. Best Practices: - Stagger outbound calls in smaller batches - Allow time between batches for system recovery - Monitor call success rates - Keep track of active calls If you're experiencing circuit overload: 1. Reduce concurrent call volume 2. Space out call attempts
2 Views