Surprisingly, the Microsoft Office 365 status website said there were no known issues.

But the Twitter account for Microsoft 365 said it had been determined that a subset of the domain controller infrastructure had become unresponsive, with the result that user connections were timing out.

We’ve determined that a subset of Domain Controller infrastructure is unresponsive, resulting in user connection time outs. We’re applying steps to mitigate the issue. More details can be found in the admin center published under EX172491.

— Microsoft 365 Status (@MSFT365Status) January 24, 2019

In November last year, Microsoft had issues with the Office 365 service, with customers complaining they were unable to sign in to their accounts.

A number of cloud services offered by the Redmond software giant use the MFA offered by Azure Active Directory to authenticate by default.

This is still down, 13 hours and counting. O365 Mailboxes in North Europe region. Apparently a domain controller is down, per MS support.

— Cybersecurity Beaumont (@GossiTheDog) January 24, 2019

And a week prior, Microsoft cloud customers using multi-factor authentication found themselves locked out of their accounts.

Office 365's support team having a lovely time in the office. pic.twitter.com/dfRV2blDJj

— Gregg Bayes-Brown (@GreggBayesBrown) January 24, 2019

Additionally, a lightning strike near one of Microsoft's South Central US data centres in September led to a massive outage in regions around the globe.

Well.. at this rate it's going to be Microsoft350 by the end of the year.

— Taneli Kaivola (@dist) January 24, 2019

The Azure service was hit by a fire suppression incident in 2017. It also suffered prolonged downtime in 2013.

Fixed your logo for you #Office365 #Office365down #Office364 pic.twitter.com/2DdUmpcn65

— Luke Pollard (@lukepollard18) January 24, 2019

@backlon @NeowinMicrosoft @thurrott This outage has prevented our business from sending and accessing email for almost an entire day, and many other businesses across Europe. Shocked by suggestions it’s a single point of failure with no failover after 6 hours here

— Adam (@stonkcat) January 24, 2019