Hi LiveKit Support, We're experiencing an issue with the LiveKit Cloud dashboard for our project starting today. Issue: The Overview dashboard is stuck in an infinite loading state — data fail to render and remain blank indefinitely. Context: - We are using LiveKit Cloud (not self-hosted) — our agents are hosted on your infrastructure - Region: We are based in France (EU/Paris region) - Issue started: Today, March 30, 2025 — no changes were made on our end Impact: We are unable to monitor our Voice AI agents and sessions Could you confirm whether there is an ongoing incident affecting the dashboard, particularly for EU/France region users? And if not, could you help us investigate the root cause? Thank you

URGENT : our livekit cloud connections are timing out with 504 errors AND we are seeing huge latency (30 seconds) between when we emit a sentence and the sentence actually being delivered over the phone through SIP). Everything indicate some issues with our livekit cloud instance. Can you please help? Email: aric@daigest.app Error we are seeing:

fetch failed: Connect Timeout Error (attempted addresses: 161.115.161.158:443, 161.115.161.153:443, timeout: 10000ms)

livekit_api::signal_client:161:livekit_api::signal_client - unexpected signal error: ws failure: HTTP error: 504 Gateway Timeout

Hi LiveKit Support, We’re evaluating LiveKit for enterprise use and want to streamline IT/security approvals. Do you provide: • Standard security/IT approval documentation or review packages • Guidance on firewall/network requirements (domains, ports, IP ranges) • Best practices or support for configuring E2EE (including grouped encryption at scale) • Options to engage directly with customer IT/security teams during review We’re aiming to reduce friction and speed up enterprise deployment approvals. Thanks, Jim

I've paid for ship plan and my account still showing build I have the invoice and everything accout: accounts@aburady.com invoice num: K5IX5T3D-0002 last 4-digits: 7897 date of payement: today 02/04/2026 Payment confirmation yes recieved from stripe

can somebody please help to investigate that?

Hi - curious if there are any plans to add additional regions to Livekit Cloud? Our primary infrastructure is in

us-west

, so we’d love to have our agent colocated nearby.

Hi, I've been using cloud livekit for a couple months now and it's great product in general. But I have two problems: • Slow Deployments: Whenever a new version of the agent is deployed it takes a considerably high amount of time of this version to be actually live and the former version stays in "Draining" state for a while. I don't know if I'm missing something or if I can setup sth to turn new version live faster. • High memory usage: I can see logs in agents instances that alerts of "process memory usage is high". I wonder if this could be the reason sometimes the agent gets stalled or gets a choppy voice

[Cloud][Upgrade] Hello*!* Is there any possibility of upgrading the machine used to run my code on LiveKit Cloud? I think I need a machine with a better processor than the current one, since I’ve been having freezing issues with my application, which is a bit heavy. It seems that increasing the RAM is not what’s needed, but rather a better processor.

I noticed this morning that my rooms are not getting deleted automatically when there are zero participants. They are still active after hours. Any ideas?

┌─────────────────┬──────────────────────┬──────────────┬────────────┐

│ RoomID          │ Name                 │ Participants │ Publishers │

├─────────────────┼──────────────────────┼──────────────┼────────────┤

│ RM_AY5wLMdZn9Hn │ TCQAIizakG4rHKu9tp2L │ 0            │ 0          │

│ RM_YNEkpU4MDFgr │ 62XakVyAZD04XkNPDN6x │ 0            │ 0          │

│ RM_YxdZQjLcoWn9 │ jhpbNJb3ypikA0Jcfwid │ 0            │ 0          │

│ RM_5DY2933Tfg83 │ hPBhTJ9ZlTcT1phfdAE1 │ 0            │ 0          │

│ RM_hcnwB3EFTT6i │ 5KrpycsEfAuKSMRKmrWQ │ 0            │ 0          │

│ RM_QrXKGRATT67B │ iedQF72QGFUoRal0WdwO │ 0            │ 0          │

│ RM_3BALZHGGxGiP │ aPd2rRnQBaWB1PKI47UD │ 0            │ 0          │

│ RM_KD4X8s5GHNTL │ 3TzyyCHMFDPkI1953lC3 │ 0            │ 0          │

│ RM_KujPjwfYK6C6 │ JWVpWuFbeHwyoSPEGxL6 │ 0            │ 0          │

│ RM_NovxH8SzBMDT │ PhNkEHYogLqb6ZXeEKlp │ 0            │ 0          │

│ RM_fhuLzDHj2MRF │ MvcmFOzvGxluFRjSOngL │ 0            │ 0          │

└─────────────────┴──────────────────────┴──────────────┴────────────┘

Hi how do I ask for region pinning? I need my agent to be in south africa

I was looking at our billing page today and noticed that our ai coustics usage seems very high. Nearly 7k minutes from less than 800 agent session minutes. Am I reading this wrong?

hi, can anyone tell me how to increase the tool timeout while using ultrvox realtime models

Hello, our client recently reported Room connection errors on some Windows machines on Firefox and Edge. We noticed they are using a VPN (so this issue might be OS independent). Their browser logged this error:

ConnectionError: could not establish signal connection: Encountered unknown websocket error during connection

I tested this on my PC with another VPN provider and did not get this very error though saw this one:

could not establish pc connection

It is also highly browser and region dependent if this error is thrown with the VPN or not. I also noticed in the network tab that the browser constantly calls a GET request to the livekit cloud

regions

before it often fails with the error from above. If this is not "fixable" then can we somehow tell that the connection failed due to an active VPN connection? In this case we can show a more detailed error to the users so they understand they shall deactivate their VPN if possible.

we've been seeing these errors for few days now and room joining and access creation endpoints are taking 17+ seconds: _app-ca3b851950aef5f9.js:99 [RTC_CLIENT] pub conn: ice candidate error - code: 701, text: STUN host lookup received error., url: stunip 143 223 91 49.host.livekit.cloud3478 (anonymous) @ _app-ca3b851950aef5f9.js:99 emit @ _app-ca3b851950aef5f9.js:105 e.pc.onicecandidateerror @ _app-ca3b851950aef5f9.js:105Understand this error 2_app-ca3b851950aef5f9.js:99 [RTC_CLIENT] pub conn: ice candidate error - code: 701, text: TURN host lookup received error., url: turnip 143 223 91 49.host.livekit.cloud3478?transport=udp (anonymous) @ _app-ca3b851950aef5f9.js:99 emit @ _app-ca3b851950aef5f9.js:105 e.pc.onicecandidateerror @ _app-ca3b851950aef5f9.js:105Understand this error 3_app-ca3b851950aef5f9.js:99 [RTC_CLIENT] pub conn: ice candidate error - code: 701, text: Address not associated with the desired network interface., url: turnslms google cloud henrf1gr.turn.livekit.cloud443?transport=tcp (anonymous) @ _app-ca3b851950aef5f9.js:99 emit @ _app-ca3b851950aef5f9.js:105 e.pc.onicecandidateerror @ _app-ca3b851950aef5f9.js:105Understand this error 3_app-ca3b851950aef5f9.js:99 [RTC_CLIENT] pub conn: ice candidate error - code: 701, text: Address not associated with the desired network interface., url: turnsomumbai1a.turn.livekit.cloud443?transport=tcp

Hey team, running into a weird issue and hoping someone can point me in the right direction. My agent sessions eventually show up in the Sessions tab of the Cloud dashboard, but there's about a 40-minute delay between when the call ends and when it appears. Is this expected? I thought sessions were supposed to show up pretty quickly after the call wraps. Some context on my setup: • Running locally via

uv run python src/agent.py dev

livekit-agents

1.5.2,

rtc-version

1.1.5, Python 3.13 on macOS Using the newer

AgentServer

+

@server.rtc_session()

pattern • Worker registered against project (

cpi-agents (dev)

) Everything looks healthy during the call itself — STT, LLM, TTS, tool calls all fire normally, and right after the session ends my logs show:

DEBUG livekit.agents uploading session report to LiveKit Cloud
DEBUG livekit.agents finished uploading

So from the agent's side the upload seems to complete within a second or two of the call ending. But the session doesn't actually surface in the dashboard until ~40 minutes later. Is there a known processing/ingest delay on the Cloud side, or is something likely misconfigured on my end? Thanks!

Hi. I am using LiveKit Cloud, and I have a zombie session, with no participants which is is showing as 730 minutes in duration. Can this be killed off somehow? RM_zxdayPECxFTM

I used free plan on Livekit Cloud to set up agents with code. Now the credits are out "Your project has exceeded its free tier limit. Please upgrade now." and I need to upgrade. But nothing happens when i click the link upgraade or billing etc. There is nowhere to add a card or move forward. What do i do? I tried both in Safari and Chrome.

Hi LiveKit team! I'd like to request access to the WhatsApp Business Calling Connector for my LiveKit Cloud project. Project details: • Project name / slug: zapclient-6bdu7pne What I need enabled: • livekit.Connector/AcceptWhatsAppCall (inbound — receive WhatsApp calls) • • livekit.Connector/InitiateWhatsAppCall (outbound — call WhatsApp users) • Current status: I already have everything wired up on my side: • Meta WhatsApp Business Calling API is already enabled on my phone number (+55 61 9866-6607, phone_number_id 1056253430896876), with voice calls toggled ON in WABA settings. • • My backend (Supabase Edge Functions) is calling POST {LIVEKIT_URL}/twirp/livekit.Connector/AcceptWhatsAppCall with the correct camelCase payload (sdp, whatsappApiKey, whatsappCallId, whatsappPhoneNumberId, whatsappCloudApiVersion, roomName, participantIdentity) and a valid server JWT signed with my LIVEKIT_API_KEY / LIVEKIT_API_SECRET. • • The endpoint is reachable but returns:HTTP 403 — "whatsapp call is not enabled for this project" • • An operator-side CallPanel is already built (LiveKit JS SDK) ready to join the room as soon as the Connector accepts the SDP. • Use case: We're a Brazilian customer-engagement platform (ZapClient) running on top of WhatsApp Business API. We need both directions: 1. Customers calling our business number → operator answers in browser. 2. 3. Operator clicks "Call" on a chat → outbound WhatsApp call to the customer. 4. Could you please enable the WhatsApp Connector feature flag (Accept + Initiate) for the project above? Happy to provide any additional info. Thanks a lot!

how to use it buying it from aws marketplace

Hey Guys, I just bough the ship plan and I want to know, what are the advantages of it ?

This message was deleted.

This message was deleted.

Would this pricing update will somehow shown in livekit inference? https://x.com/ElevenLabs/status/2052388133585436810

Subject: Inbound SIP trunk returns "404 No trunk found" for all INVITEs — Twilio Elastic Trunk integration Project: scale-desk-ai-nxvkobrt SIP host: scale-desk-ai-nxvkobrt.sip.livekit.cloud Inbound trunk ID: ST_teYSM6uLNDbF Dispatch rule ID: SDR_sofRZG5hk4Rn Issue Every inbound SIP INVITE to my LiveKit Cloud project returns SIP/2.0 404 No trunk found, regardless of the trunk configuration. As a result, no rooms are ever created and the agent worker never receives a job. OPTIONS to the same SIP host returns 200 OK over TLS, so the edge is reachable. Configurations tested (all return 404) LiveKit inbound trunk variations: - numbers=["+16508558750"] only - auth_username + auth_password only - auth_username + auth_password + numbers - allowed_addresses=["<my public IP>/32"] (verified IP via Via: received=…) - allowed_addresses=["0.0.0.0/0"] - Full Twilio published SIP signaling CIDRs in allowed_addresses - Combinations of all of the above Twilio side variations: - <Dial><Sip> TwiML with transport=tls, transport=tcp, with/without username+password attributes - Twilio Elastic SIP Trunk with secure=true + sipshost5061 - Twilio Elastic SIP Trunk with secure=false + sip:host;transport=tcp (exactly per LiveKit's docs) - Twilio Elastic SIP Trunk with embedded credentials in origination URI (sipuserpass@host:5061;transport=tls) Reproducing 404 directly Sending a raw SIP INVITE over TLS from an external host to sip:+16508558750@scale-desk-ai-nxvkobrt.sip.livekit.cloud:5061 consistently returns: SIP/2.0 100 Processing Record-Route: sip:10.244.41.131:40004;sipfe;transport=tcp;lr Record-Route: sip:139.185.54.113:5061;sipfe;transport=tls;lr Via: SIP/2.0/TLS <client-ip>:<port>;branch=<branch>;received=<public-ip>;rport=<port> … SIP/2.0 404 No trunk found Sample Call-IDs from probes: probe-c7fd41f4e4851501@probe.local, probe-dc3200552a345e9d@probe.local. Real Twilio call SIDs failing at 0 seconds: CAf8f5a0bf25b2a6d8a30adaeae5e8d017, CA7dc67a3def476c54f23c6e3c76266df4 (timestamps around 2026-05-09 07:05 UTC). Twilio details - Account type: Full, active - Twilio trunk: TK1c2fccd6ca7df03597b180ed033d4ef4 - Domain: scaledesk-1778310147.pstn.twilio.com - Origination URL: sip:scale-desk-ai-nxvkobrt.sip.livekit.cloud;transport=tcp - Phone number: +16508558750 Question What additional matching criteria does your inbound trunk matcher require beyond the documented auth_username/auth_password, allowed_addresses, and numbers fields? The docs state that setting either auth or allowed_addresses is sufficient, but neither path matches incoming INVITEs on this project. Could you check your edge logs for the timestamps above and tell me why the matcher is returning 404 for these INVITEs?

Is there a way to get the traces from cloud hosted agents?

hi! I got this email saying I reached the limit on observations for this project but I have agent observability off for all my projects including this one AND this project hasn’t run a session ever. We created it for future use but haven’t used it at all I’ve gotten 37 of these in the past 3 days, anyone know how I can make it stop?

Hi, the pricing page says the Ship plan does not support

Metrics export APIs : Query and export analytics or telemetry data to external systems

Ok for the API, but does it mean we can't even export data opentelemetry ?

Hi Team, Wanted to flag a strange issue we're noticing in the LiveKit voice agent pipeline — when the agent speaks, some completely unrelated background noise is audible in the audio. It's not static or echo — it sounds like external sounds such as laughter, music, or other ambient noise that has no relation to the agent or the call. This seems unusual and could indicate audio from another source bleeding into the agent's audio track. Please help us here. https://cloud.livekit.io/projects/p_7hoiye4mqae/sessions/RM_BX2E4fL3AWQi/observability?mode=transcript - (02:05) https://cloud.livekit.io/projects/p_7hoiye4mqae/sessions/RM_a9wSx8BqcjuA/observability?mode=transcript - (01:59 https://cloud.livekit.io/projects/p_7hoiye4mqae/sessions/RM_TQYpph8b7qAh/observability?mode=transcript - (00:36)

archived the channel