Error starting calls
# support
r
Each time I attempt to initiate a call, the assistant speaks only one word before abruptly stopping, resulting in the call ending. The reason for the termination is displayed as an unknown error. For instance, the call IDs are 0e26c5e5-1a0d-47f8-a118-b2b26447535c and e94e342c-bd63-4ed4-aba2-3a415b7c6f0c. I have also tried this with default agents, but the issue persists. After attempting to start the call two or three times, I encounter a concurrency limit error, even though none of the calls were successful. Everything was functioning properly until a couple of days ago. Please let me know if any additional information is required from my end.
v
Hi @rajatsablok, I'm an AI assistant for Vapi. I'll analyze available sources and get back to you shortly.
----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Call Ended Reason]() * [List Calls]() * [Create Phone Call]() * [Get Call]() -----------------------------
### Answer It seems you're encountering a few issues with starting calls on Vapi, specifically with abrupt terminations described as "unknown errors" and hitting concurrency limits. Here are some steps and considerations that might help resolve these issues: 1. **Unknown Error**: The "unknown error" during calls can be particularly challenging as it does not provide specific information about what went wrong. I recommend checking the server-side logs for any additional error messages or unusual activity around the time the errors occurred. 2. **Concurrency Limit Error**: This error typically indicates that multiple call attempts are being made simultaneously beyond the limit allowed by your current Vapi plan or system setup. Ensure that any retries or parallel call setups are managed to stay within the concurrency limits. Adjusting the retry logic and ensuring proper call teardown can help manage this limit more effectively. 3. **Debugging Steps**: - Review the implementation of the call setup in your code to ensure that it aligns with Vapi’s best practices, particularly around error handling and concurrency management. - Test the call flow with minimal setup to isolate whether the issue is with specific configurations or broader system-level problems. 4. **Consult Documentation and Support**: - Ensure that your system’s implementation aligns with the guidelines and configurations suggested in the Vapi documentation. You might find sections like error handling, call setup, and concurrency management particularly useful. - If the problem persists and you're unable to identify the cause from logs and system checks, I recommend reaching out to Vapi support with detailed information including the call IDs, error messages, and descriptions of the steps already taken to troubleshoot the issue. By methodically checking these areas, you can better iso
late the cause of the issues and apply targeted fixes. If there's any other information or specific logs you can provide, that might also help in diagnosing the problem more accurately.
r
I have not implemented any server-side logic with Vapi till now and am only testing via dashboard. But I facing issues validating my use case with the dashboard itself.
v
Got it. Tagging @User for additional help.
s
Can you once try to use whisper and lmk whether this fix this issue or not?
r
Worked, I think the issue was with the Deepgram API keys that I added Thanks for the help! Also, is it possible to include such things in the ended_reason of call?
s
Yup, remove the key and it will work.
You can use end-call-report for that purpose