VAPI calls stuck as in-progress — finalization pipeline severely delayed We're experiencing an issue where web calls remain stuck as in-progress for 2+ hours after ending. This started around 10:30 UTC today (March 21). Evidence that this is server-side: - The Live Call Control API returns "Call Not Active" for these calls, confirming VAPI's real-time infrastructure ended them — but the REST API still shows status: "in-progress", cost: 0, no endedAt - Same assistant that worked fine yesterday is affected today with zero config changes - Happens regardless of webhook server URL (tested with both dev tunnels and Hookdeck) - Calls DO eventually finalize after ~2 hours instead of the usual 30-120 seconds - Deleting stuck calls via the API sets endedReason: "call-deleted" but status stays in-progress permanently — the finalization pipeline never picks them up This blocks all post-call processing since end-of-call-report webhooks are delayed by the same amount. Also, ghost calls consume concurrency slots. Is anyone else seeing this, or is this account-specific? Any insight into what's happening with the finalization pipeline today?

Hi there, Thank you for your message. Our team is currently out of the office. We operate Monday through Friday, from 9:00 AM to 8:00 PM Pacific Standard Time (PST). We’ll get back to you as soon as possible during our normal business hours. If your message is urgent, please mark it accordingly or include “URGENT” in the subject line, and we’ll do our best to respond promptly. Warm regards, Vapi Customer Support Team

Action Required: To escalate this critical infrastructure issue, submit a detailed report to Vapi support including your call IDs, timestamps, and a description of the problem. Use the following format for clarity:

Call ID: [your_call_id]
Timestamp: [when issue occurred]
Issue: Calls stuck as in-progress for 2+ hours after ending
Expected: Calls should finalize within 30-120 seconds

Submit this via the [Vapi issue reporting page](https://docs.vapi.ai/issue-reporting) for fastest resolution. Source: - [Vapi Issue Reporting Guide](https://docs.vapi.ai/issue-reporting)

Call ID: 019d1027-691a-7000-9f5c-d8fe3b68c118 Timestamp: 2026-03-21 11:28 UTC Issue: Call stuck as in-progress for 2+ hours after ending. Live Call Control API confirms call is "Not Active" but REST API still shows in-progress, cost: 0, no endedAt. end-of-call-report webhook never fired. Expected: Call should finalize within 30-120 seconds. Call ID: 019d1016-8d17-7000-939c-06a1d8935831 Timestamp: 2026-03-21 11:10 UTC Issue: Same as above. Call was deleted via API but status remains in-progress with endedReason: call-deleted — finalization never ran. Expected: Delete should transition status to ended. Additional context: - Org ID: 0ed9d25a-9d8f-49f4-b101-64aea17ec830 - 6 calls affected since ~10:30 UTC today - 2 eventually recovered on their own after ~2 hours - 4 that were deleted via API are permanently stuck - Tested with both dev tunnel and Hookdeck webhook URLs — same result - Same assistants worked fine yesterday with zero config changes

This definitely looks like a backend finalization pipeline delay, not anything on your config — especially since Live Call Control shows “not active” but REST never transitions. The 2+ hour lag + missed end-of-call webhooks + stuck concurrency is the real blocker here. The fact that deleted calls never finalize suggests the cleanup queue is failing or backlogged. Are you seeing this across all regions or tied to a specific provider/region (e.g. Telnyx, WebRTC)? Also, have you noticed if transcripts/recordings are delayed too, or just the status + webhook layer? @SwordFab - Lumara

Guys still hapening, we are having 2h and more delay on any of our calls, this is unacceptable, we used the status update + ended status to at least close them on our be, but we need the call url that comes from the end-of-call-report, we are forced to migrate to 11labs agents

@SwordFab - Lumara sorry about the trouble you're experiencing. could you provide us the payload for the call request if you are using the API to start calls?

Sorry, I don't have the bandwidth to help debug this further. I've already migrated to ElevenLabs as your API has been quite unstable lately. I’ve gone ahead and fixed the ticket myself since it was urgent. You should investigate what happens when the upload bucket is deleted; I suspect your jobs are failing, the exponential backoff reaches its limit and allows the finalization pipeline to proceed regardless. But you guys should return an error about it, cmon…