Genesys Cloud - Main

 View Only

Sign Up

Expand all | Collapse all

Call Error message

  • 1.  Call Error message

    Posted 15 days ago

    Hi all,

    Today, a few of our users have been getting sporadic messages indicating WebRTC endpoint disconnect, reason=iceIdleDetection. Anyone seen this before or know what its in reference to?


    #Omni-ChannelDesktop/UserInterface
    #Unsure/Other

    ------------------------------
    John Butine
    ------------------------------


  • 2.  RE: Call Error message

    Posted 15 days ago

    Olá @John Butine,

    Sim - a mensagem iceIdleDetection está relacionada ao mecanismo ICE (Interactive Connectivity Establishment) do WebRTC.
    Na prática, significa que a sessão de mídia parou de receber pacotes de keep-alive STUN/RTCP, e o endpoint interpretou a conexão como "inativa".

    Isso costuma acontecer quando:

    • Há perda de pacotes ou instabilidade no Wi-Fi
      Redes domésticas com oscilação ou roteadores que entram em modo de economia de energia podem interromper o fluxo.

    • Firewall / proxy / VPN encerra sessões UDP por tempo ocioso
      Muitos equipamentos corporativos têm timers de NAT agressivos, encerrando a conexão após 30–60s sem tráfego.

    • O usuário troca de rede durante a chamada
      Ex.: Wi-Fi → 4G, ambiente corporativo → residencial, troca de AP no mesmo prédio.

    • O navegador ou o sistema operacional suspende a aba/processo
      Isso pode ocorrer por políticas de economia de bateria ou gerenciamento de recursos do Chrome/Edge.

    Quando o ICE deixa de receber respostas de keep-alive, ele entende que o caminho de mídia morreu e dispara o iceIdleDetection.

    O que verificar imediatamente

    • Se os usuários impactados estão na mesma rede, no mesmo Wi-Fi, no mesmo roteador ou na mesma VPN.

    • Se há NAT timeout ou inspeção de tráfego ocorrendo no firewall/SWG/VPN.

    • Teste com conexão cabeada para comparar com o Wi-Fi.

    • Veja se o navegador está com limitação de energia/bateria ativa (Windows/Chrome/Linux/Mac).


    No contexto do Genesys Cloud

    • Garanta que a rede atenda aos requisitos de WebRTC:
      UDP 16384–32768 liberado, STUN sem inspeção ou reescrita.

    • Certifique-se de que nenhum firewall ou proxy esteja bloqueando ou modificando pacotes STUN.

    • Verifique se o navegador não está sendo throttled (gerenciamento agressivo de energia do laptop pode causar isso).



    ------------------------------
    Fernando Sotto dos Santos
    Consultor Grupo Casas Bahia
    ------------------------------



  • 3.  RE: Call Error message

    Posted 15 days ago

    Good deal, we will monitor this, thanks to you both @Luiz Rosa @Fernando Sotto dos Santos for the quick replies!



    ------------------------------
    John Butine
    ------------------------------



  • 4.  RE: Call Error message
    Best Answer

    Posted 15 days ago
    Edited by Jason Kleitz 14 days ago

    Hi John Butine,

    This happened to me recently, and in my case, it was related to the agent's network.

    This message appears when Genesys Cloud stops receiving RTP packets from the agent's station for 10 seconds, triggering ICE idle detection and disconnecting the call.

    More details: https://developer.genesys.cloud/platform/operational-event-catalog/telephony/telephony-0010

    I recommend collecting the browser logs when the error occurs.

    If the issue persists, run the WebRTC diagnostic tool after the issue is encountered, and if it continues, open a Genesys Cloud support case for further analysis.


    ------------------------------
    Luiz Rosa
    Full stack developer
    ------------------------------



  • 5.  RE: Call Error message

    Posted 14 days ago

    Can you help me understand how, from the event, I'm able to identify which user was impacted? 

    I'm not sure what the entityId refers to - it's not a user and not a phone. I can see in the additional event details there is a conversationId, but what if there are multiple agents on that conversation? Do I need to query each the conversation each time and check which agent(s) have an error in the disconnectReason?

    I'm looking at integrating this with event bridge to build a more useful table of these events, so I can try to build up a picture about which agent(s) have it re-occurring frequently. Why that info isn't available in the Event itself, I'm not sure.



    ------------------------------
    James Dunn
    Telecoms Specialist
    ------------------------------



  • 6.  RE: Call Error message

    Posted 13 days ago

    Hi James Dunn

    In the conversation history the affected participant shows disconnectType = ERROR, confirming that the disconnection occurred on the agent's side.

    Example: 

    API:

    api/v2/analytics/conversations/details/query
    In the segments of this endpoint, it's possible to retrieve the participant's error code: 
      "segments": [
      {
        "conference": false,
        "queueId": "1ef6bb14-a645-419b-8602-5a2751db4a6c",
        "segmentStart": "2025-10-24T14:50:29.612Z",
        "segmentEnd": "2025-10-24T14:50:29.722Z",
        "segmentType": "alert"
      },
      {
        "conference": false,
        "disconnectType": "ERROR",
        "errorCode": "error.ininedgecontrol.connection.webrtc.endpoint.disconnect.iceIdleDetection",
        "queueId": "1ef6bb14-a645-419b-8602-5a2751db4a6c",
        "segmentStart": "2025-10-24T14:50:29.722Z",
        "segmentEnd": "2025-10-24T15:01:01.187Z",
        "segmentType": "interact"
      },
      {
        "conference": false,
        "disconnectType": "ERROR",
        "errorCode": "error.ininedgecontrol.connection.webrtc.endpoint.disconnect.iceIdleDetection",
        "queueId": "1ef6bb14-a645-419b-8602-5a2751db4a6c",
        "segmentStart": "2025-10-24T15:01:01.187Z",
        "segmentEnd": "2025-10-24T15:01:46.187Z",
        "segmentType": "wrapup",
        "wrapUpCode": "ININ-WRAP-UP-TIMEOUT"
      }
    ]
    It's also possible to add this metric to the interface and view it.

    Holpe this helps.



    ------------------------------
    Luiz Rosa
    Full stack developer
    ------------------------------



  • 7.  RE: Call Error message

    Posted 14 days ago
    Edited by Steve Hennesy 14 days ago

    Have you or anyone else successfully resolved this issue?  
    ------------------------------
    Steve Hennesy
    IT
    ------------------------------



  • 8.  RE: Call Error message

    Posted 13 days ago

    Hi Steve Hennesy

    In our case, we ran the WebRTC Diagnostic Tool and it was clear that the issue was related to the network at the agent's location. We then forwarded the case to the customer's network team.



    ------------------------------
    Luiz Rosa
    Full stack developer
    ------------------------------



  • 9.  RE: Call Error message

    Posted 10 days ago

    I was out for a few days, but it seems like the issues have resolved, no further reports of the disconnection error messages. I would imagine it was related to VPN or onsite network blips for isolated users, not the org as a whole. Will continue monitoring and update here as needed. Thanks @Luiz Rosa for the insight here!



    ------------------------------
    John Butine
    ------------------------------



  • 10.  RE: Call Error message

    Posted 15 days ago

    Hi all,

    We are experiencing this problem in 6 Genesys organizations located in 3 different AWS zones (us-east-1, us-west-1, and sa-east-1), with agent operations in three countries: Brazil, Argentina, and Mexico.

    Calls are being disconnected with the error error.ininedgecontrol.connection.webrtc.endpoint.disconnect.iceIdleDetection, and in many situations the agent's screen freezes, preventing them from performing any action, such as trying to click the button to end the call that has already been disconnected due to the failure.



    ------------------------------
    Gunter Fischborn
    Voice Specialist
    ------------------------------



  • 11.  RE: Call Error message

    Posted 13 days ago

    Hi Gunter Fischborn,

    I'm from Brazil and work with both regions, us-east-1 and sa-east-1. In our cases, we identified that the issue was on the customer's side. I recommend collecting the browser logs when the error occurs. If the issue persists, run the WebRTC Diagnostic Tool right after the problem happens.

    If it continues, please open a Genesys Cloud support case for further analysis.



    ------------------------------
    Luiz Rosa
    Full stack developer
    ------------------------------