Azure's wobbly day as three services glitch around the world
Un-planned PITSTOPs for Visual Studio Team Services, App Services and DocumentDB
If your favourite Microsoft staffer isn't answering the phone today, it may be because they're getting calls from angry Azure workers all over the world after the cloud service experienced three Partial Inability To Support Totally Optimal Performance (PITSTOP) incidents.
DocumentDB was the first to wobble, producing a “No data available” error when users tried to view Monitoring Metrics for Total Requests. Not a killer, but an annoyance. Microsoft says it has a fix and will advise when it's been made, but is yet to do so 14 hours after notifying users of the glitch.
Next came a problem for App Service \ Web Apps and Virtual Machines users in the in East Asia region. That problem later meant higher-than-usual latency for those trying to reach cloudy VMs. Microsoft's pinned this one on “an interruption to our physical network infrastructure within the region” and is hustling to get things fixed.
The third PITSTOP is impacting Visual Studio Team Services and started in West Europe but later spread to Australia East, East US, East US 2, North Central US, South Central US too. This problem was a double-header: first a “collection database” went flaky, then it became impossible to add new users. Both problems have since been resolved.
Azure's a pretty reliable cloud, but three wobbles in one day is un-nerving. There's no word on whether the three are in any way related (the infrastructure outage suggests not), but Microsoft or any other cloud operator can do without clumped PITSTOPs like this. ®