My Brisbane-based colleague Glen recently reported an interesting problem he’d encountered with calls to O365 UM going unanswered, and in some cases bouncing back to the Attendant.
In the FE he was seeing these diagnostic messages:
ms-diagnostics : 10049;source="LYNCFE01.contoso.com";reason="The current signaling session took too long to establish the connection, so the call leg was canceled.";component="MediationServer"
He also noticed a pattern: all the calls that failed were being routed to a server(s) in Hong Kong (e.g. HK20C00EDG02.LYNC0C001.LOCAL), where those that succeeded were going elsewhere.
He logged it with O365 Support and they were promptly able to resolve the issue, although they didn’t reveal the root cause.