Microsoft suffered a service disruption over the weekend after a power incident at an Azure datacenter in the West US region affected Windows Update.
According to the Windows giant, between 0800 UTC on February 7 and 0423 UTC on February 8, “multiple Azure services” in the region suffered “intermittent service unavailability, timeouts, or increased latency.”
A wide variety of services were affected, including the Azure Kubernetes Service, Azure Confidential Compute and, ironically, Azure Site Recovery.
Microsoft blamed the stoppage on “an unexpected interruption to utility power at one of our West US datacenters,” which, “resulted in a loss of power to parts of the facility.” Backup power systems were activated, but it took until 0423 UTC on February 8 before Microsoft declared the issue mitigated.
However, rather than just being another outage in a long line of problems, this one hit Windows Update and the Microsoft Store. At 1634 UTC on February 7, Microsoft admitted that the datacenter power outage had also impacted “customers’ ability to complete operations in the Microsoft Store and Windows Update.”
“As a result, some users may experience failures or timeouts when installing or updating Microsoft Store apps, or when downloading Windows updates.”
Affected administrators received almost a three-hour reprieve from the company’s update flow while Microsoft’s engineers worked on the problem. By 1930 UTC on February 7, the company confirmed that the services had been restored. It added: “Some IT administrators may continue to experience residual latency affecting Windows Server update operations, including the Windows Server sync API. This Server-specific issue is being addressed separately.”
Microsoft has had a bad run of incidents and outages lately. As well as having to push out out-of-band releases to deal with broken updates, the Azure cloud service has had more than its fair share of problems. Blaming a degradation of service in the West US region on an “unexpected interruption to utility power” is likely to raise questions, given the widespread use of UPS systems and the time taken to declare the incident mitigated.
Alternatively, with CEO Satya Nadella deciding that Microsoft needs someone to focus on engineering quality, perhaps the first act of the appointee was to choke off Windows Update. Certainly, losing several Azure services in the West US region is less than ideal, but at least there were a few hours without Windows Updates. ®