Is there another supabase outage? This happened last week or the week before - my calls all of a sudden are not detecting voicemails and timing out rather than hanging up. Example call ids: 231178c2-1ca8-430a-8384-86038a2ab272, e774ce17-f6e1-416c-b578-725a6606b739, fb431482-062d-4042-8e65-74d650fd2359.

@jsieber-eip I'm sorry to hear you're experiencing issues with your calls. To assist you better, I'll need to check the status of the calls you've mentioned. However, as a support bot for VAPI, I don't have direct access to external systems like Supabase or the ability to check call statuses. If you suspect there's an issue with the VAPI service itself, such as a problem with detecting voicemails or calls timing out, I recommend checking the status page for any reported outages or incidents. Additionally, you can review your implementation to ensure that the voicemail detection settings are configured correctly. If you've already done this and the problem persists, please reach out to support with the call IDs you've provided for a more in-depth investigation. They will be able to check the logs and provide you with specific information regarding the calls in question. If there's anything else I can assist you with, please let me know!

Just happened again at about the same time, is this when supabase rolls out their releases? or when the vapi team rolls out their releases?

@jsieber-eip Set **startSpeakingPlan.waitSeconds **to 0.8 and transcriber.endpointing to 300 for deepgram and then make the calls.

That should not be the case - it randomly stopped @Shubham Bajaj it will not detect any voice mails for about an hour, then it begins working again. It happens for time periods, randomly, thens tarts working again. A setting issue would be deterministically failing, not randomly failing for periods of time . It has happend 3x so far, and each time it is not an isolated issue, just a few hours of failures.

Hey Jsieber-sep, you can give it a try for few test calls, and eventually you will notice the difference. Allowing Assistant to wait more will help you because its related to endpointing, and sometimes AI speech Models can't find out if user has finished speaking correctly.

It is working consistently right now... it only randomly stops working for periods of time. That's not a settings issue? unless there is additional latency randomly from the voicemail detection? @Shubham Bajaj

No outrage has nothing to do with it. It's an endpointing related thing where sometimes speech models cant produce the expected results in required time and on top of it you have twilio voicemail detection. These systems are not perfect so sometimes few hiccups can happen. @jsieber-eip Let me know if it happens for you again.

okay. so you're saying there is a random latency bump for a period of time, maybe due to a networking issue and thats it? that would make sense

> okay. so you're saying there is a random latency bump for a period of time, maybe due to a networking issue and thats it? that would make sense It's not related to latency rather endpointing, which decides when user has finished speaking based on audio and transcription nothing to do with latency instead model judgment using the config failed for the shared call ID. Happy to chat more on endpointing if still unclear to you.

i will share more on this thread if we have another one of the periodic outages. when it occurs, it stops detecting vm at all for an hour or two , and then starts working again later. it has not occurred since i started the thread. It's not just a few hiccups though, its periods of time where zero voicemail detection occurs, and then no issues outside of that. You can reference the call ids in my first post if you want an example, every call within their time range did not detect voicemail.

Adding this thread to my notes list.