Microsoft's Windows Azure cloud experienced a substantial outage on Saturday, with key services in the Western United States region being affected for over nine hours.
According to the Azure service dashboard, the problem started at 0300 UTC (1400 AEDT).
"We are experiencing an issue impacting availability of Compute, Service Bus and Web Sites Services in the West US sub-region," Microsoft said.
"We are actively investigating this issue and working to resolve it as soon as possible. Further updates will be published to keep you apprised of the situation.
"We apologise for any inconvenience this causes our customers."
At 0600 UTC Azure support staff said they had identified the root cause of the problem - "a networking issue" - and at 1216 UTC all services were declared restored:
"The repair steps have been executed successfully, the availability of the services hosted in the cluster has been restored and verified," an Azure notification stated.
An Australian Azure customer said they were affected by the outage and had enacted disaster recovery plans, shifting to Azure's US East Coast region within 30 minutes of detecting the problem.
While the customer expressed disappointment that Azure went down at all, they added that "if we were on physical infrastructure, things today would have looked very bad for a very long time."
"We did not think we would do a live disaster recover test so soon," the customer said.