Azure's Australia East region, located in Sydney, experienced nearly eight hours of degraded performance thanks to what Microsoft is calling “an ongoing Networking issue.”
Microsoft's incident report says “Between 19:10 UTC on 20 Sep, 2016 and 02;45 UTC on 21 Sep, 2016, some Azure Services in Australia East region may have experienced intermittent connectivity issue to their services, or experience errors when performing Service Management operations.”
Australia is 10 hours ahead of UTC, so the problem persisted through most of Wednesday morning in Australia, a busy time when business would have been keen to have Azure working.
Whatever the time of the incident, it meant Azure's Stream analytics and media services, IoT hub, SQL databases, Web apps, Logic apps, HDinsight, Visual Studio team services and API management.
In an unpleasant irony, even Azure's Site Recovery disaster recovery offering had problems: the advisory says “Users may have observed delays with VM replication from on-premises to Azure. Users may have also experienced delays with Azure Site Recovery operations from the portal.”
Microsoft says the underlying cause of the brownout was “a connectivity issue within the Australia East”. The glitch was isolated “to fully mitigate this issue.” Which sounds like a fat data pipe into a Sydney bit barn failed badly – perhaps due to backhoe action - and had to be routed around.
If you know more, feel free to let us know. ®