Credit: Dreamstime
An authentication problem is impacting some users globally that are attempting to access their Microsoft 365 services.
Dynamics 365, LinkedIn, and other services that leverage Azure Active Directory (AAD) are also impacted, Microsoft said in a status update on the issue, which has spread across Australia, UK, Europe and US.
“Customers that have their authorisation cached are unaffected by this issue, and new authentications are succeeded approximately 50 percent of the time and rising,” Microsoft said.
Microsoft noted the issue arose on Tuesday, January 29 at 9:15 PM UTC (AEST: Wednesday, January 30 7.15am).
It noted the preliminary root cause of the problem stemmed from a portion of a third-party managed network infrastructure that facilitates authentication requests is degraded, which impacted access to one or more services.
According to Microsoft, services have been moved to an alternate network provider and it has performed numerous targeted service restarts.
“Our telemetry indicates that service availability has improved significantly and most authentication requests are succeeding; however, we still are observing residual impact to some services," the statement said.
"We're working to develop a remediation strategy for all remaining impact."
In November, Microsoft Office 365 users faced two outages. The first occurred on 19 November, blocking users from logging into their accounts using Multi-Factor Authentication (MFA).
The second outage took place eight days later on 27 November.
At the time, Microsoft said customers using Multi-Factor Authentication (MFA) may experience intermittent issues signing into Azure resources, such as Azure Active Directory, when MFA is required by policy.
Two and a half hours later, the dashboard reported that after resolving a problem with an earlier DNS (Domain Name Service) issue, engineers rebooted the services.
Microsoft also released a detailed report into its 14-hour failure on 19 November where it had identified three root causes, detailed the failures and steps engineers took to recover the service, and steps it plans to take over the next two-plus months to review and update its processes and procedures.
