Anyone else seeing intermittent call failures on weekly2 since ~Apr 8? Since 2026-04-08 ~16:15 UTC, inbound calls intermittently: short ring → ~9s silence → hangup. Twilio side: error 31901 WebSocket Connection Timeout. Twilio sends the WS upgrade to

phone-call-websocket.aws-us-west-2-backend-production-weekly2.vapi.ai/{callId}/transport

and gets no response. SIP is clean (

200 OK

, PDD 102ms) — purely the media handshake fails. VAPI dashboard for affected calls: -

ended_reason: customer-did-not-answer

-

messages_count: 0

- no recording/transcript - Start Time = N/A - end-of-call-report arrives ~6 min later. - Create-call itself returns

200 queued

with a valid call ID — failure is downstream of that. Looks like infra, not config: - Intermittent, same config/assistant — some calls work, some don't - Much worse under concurrency - Affects multiple unrelated accounts - Started cleanly at one timestamp, no deploy on our side - Our numbers were moved weekly1 → weekly2 morning of Apr 7, ~36h before failures started Tried: - Switched to Daily → rapid-fire tests all worked - Switched back to weekly2 late Fri night → also worked - Can't tell if it's fixed or just quiet Fri night Questions: 1. Anyone else on weekly2 / Weekly channel seeing WS handshake timeouts the last few days? 2. Anyone found a way to get moved to / stay off a specific weekly cluster? 3. Anyone noticed weekly2 get better in the last ~18h? Customer demo Monday on this infra, support ticket open, mostly hoping someone recognizes this. 🙏