Microsoft cloud users across the globe this week moaned that they were not able to connect with a plethora of services for several hours due to the DNS borkage.
A range of Microsoft services was impacted, together with Azure databases, Microsoft 365, Teams, Dynamics, OneDrive, SharePoint Online and Compute offerings.
Following reviews of the flaw, Microsoft posted an Azure popularity web page replace stating: “Customers may revel in intermittent connectivity problems with Azure and other Microsoft services (such as M365, Dynamics, DevOps, and many others).
“Engineers are investigating DNS decision problems affecting network connectivity. Connectivity troubles are resulting in downstream effect to Compute, Storage, and Database offerings, and a few customers may be unable to report guide requests.
“More statistics can be provided as it will become available. Some customers might also start to see healing.”
After investigating the outage, Microsoft confirmed that “users can be unable to get admission to Microsoft 365 services or functions”, adding that it had “recognized and corrected a DNS configuration problem that prevented customers from gaining access to Microsoft 365 offerings and features”
“We’ve determined an increase in success connections and our telemetry suggests that all services are recuperating. We’re continuing to reveal the environment to validate that provider has been restored.”
The three-hour outage has because come to an give up, with Microsoft confirming that its engineers have mitigated the issue and those maximum offerings were recovered.
“Engineers diagnosed the underlying root cause as a nameserver delegation change affecting DNS decision and resulting in a downstream impact to Compute, Storage, App Service, AAD, and SQL Database offerings,” it stated.
“During the migration of a legacy DNS gadget to Azure DNS, a few domains for Microsoft offerings were incorrectly up to date. No consumer DNS facts had been impacted throughout this incident, and the supply of Azure DNS remained at one hundred% throughout the incident. The hassle impacted the best facts for Microsoft services.
“To mitigate, engineers corrected the nameserver delegation trouble. Applications and services that accessed the incorrectly configured domain names may additionally have cached the wrong statistics, main to a longer recuperation time until their cached data expired.” µ