Amazon Connect CCP - Softphone call failed, WebRTC issue, Red screen of death

4

We use the Amazon Connect CCP in our call center with around 300 agents. Agents frequently report errors on the CCP during both inbound calls and outbound calls. The errors are typically "Softphone call failed: Failed to establish softphone connection" or "WebRTC issue: Call failed due to a browser-side WebRTC issue" displayed on the CCP with red background. Most of these agents report encountering these errors occasionally i.e they are able to answer calls and make calls normally during their shift but suddenly these errors occur on the next few calls. Typically the agent then logs out of Connect , restarts the browser and/or the computer and then is able to receive calls successfully until maybe a few hours later when these errors resurface. The opinion out there seems to be these are due to network issues related to internet connectivity, however there are no connection issues with any other applications or webpages at the time when these softphone errors occur, only the CCP softphone. We have never gotten a satisfactory answer from AWS Support and constantly having to restart/relogin is not a long-term solution. This same issue has been raised by other users on the connect streams Github but with no solution : https://github.com/amazon-connect/amazon-connect-streams/issues/428. I'm curious to hear if any other Amazon Connect users have resolved this satisfactorily.

asked 2 years ago8823 views
2 Answers
5

This is more common than you would think when it comes to cloud contact centers, specially with the WFH phenomenon. What I would do is put some sort of monitoring agent on the agent desktop and see what kind of network stats you get from there. Remember that voice is very very sensitive to network issues so that might be the most likely reason why you’re not seeing it manifest anywhere else. There’s a company out of Australia that has a cool addon over CCP that handles this use case. You might want to give them a shout too https://operata.com/

david

profile picture
dmacias
answered 2 years ago
  • Seconding the suggestion of Operata. We've got a client that uses it and with working from home it has been invaluable in tracking down these issues. It not only tracks network issues but also now tracks CPU and memory use of the agents' computer. It can also alert agents to issues before they become noticeable.

  • Thank you for your response. Many of our agents are WFH which of course puts their network connectivity outside our control.

  • The connectivity, even if outside your control can be very valuable to understand in order to improve/investigate any issues. Network performance, ISP, Connectivity Method (Wi-fi/Wired), VPN Connections, etc all contribute to your agents and customer experience.

    Understanding these patterns, especially over time, is likely key to the root cause of your issues. E.g. we have seen instances where agents have perfect networks until 3:30 each day (when their kids returned home), at which point quality and error rates would suddenly spring up. Once understood, you can then work around these issues

1

For transparency, I am from Operata (thanks for the kind words David and ledge)

I would echo the comments from David and ledge, that Voice connectivity, especially when browser-based, can have different demands and constraints than other applications on the same network. Amazon Connect is great at mitigating many of the issues that can impact voice, however even it has limitations.

Added to that, there is a range of other elements beyond the network that can impact your agents - Equipment, Preheperals, Security/Proxy/Firewalls, Software, Connectivity, Location, Browser, Behavior, and much more. Proactive monitoring is your best tool to understand which of these elements could be impacting your service as a whole and which ones could be contributing factors to any specific issues raised.

If you can provide any more details on your specific scenario, I will provide any insight I can, additionally, I am also happy to walk you through Operata, with a focus on your issue and what the platform can do to help identify the root cause.

answered 2 years ago
  • Thank you for your response. These errors occur even before the voice data has begun, i.e the agent hasn't even answered the call yet, it has either just started ringing or the agent has just hit Accept which means the webRTC connection is just being established when the error occurs.
    Most of our agents are WFH, and they either use their own devices which may be Windows/Mac/Linux or a company provided computer which are all Ubuntu Linux. We enforce minimum requirements for agents in terms of hardware , internet speed and browser versions. Generally, things seem to be fine (which I think eliminates any Firewall issue) but the occurrence of these errors is more frequent than we would like.

  • I would love to know more about your environment, your agent's environments, and the agents experiencing issues. Our experience suggests when not enforced, not all agents follow the guidelines, to their own detriment - sometimes intentionally.

    Re: firewalls, issues arise from VPN, routing, or security permissions. E.g. Security teams whitelist Amazon Connect by IP, which has not been updated as Amazon releases new servers.

    While happy to continue to comment, Operata also has a zero-cost review if you would like us to take a deeper look. https://operata.com/blog/accelerate-by-operata

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions