The honest view from inside technical leadership
Analysis

Engineering leadership

Independent reporting on the decisions, costs and contracts that shape how engineering organizations build with AI.

Leadership · News

Microsoft 365 Outage Ran 30 Hours After a Network Capacity Upgrade Went Wrong

The change that broke Outlook and Teams in more than 70 countries was meant to add network capacity. Microsoft brought services back in stages, and its final all-clear came more than 30 hours after the start time it now lists.

Line-art illustration of a network switch with an upgrade arrow and a queue of envelopes waiting behind a gate, representing the Microsoft 365 outage caused by a network capacity upgrade

The Microsoft 365 outage that began on Thursday, 12 November 2026 lasted about 30 hours, according to Microsoft’s service health page, which lists a start time of 10:05 UTC that day and an end time of 16:10 UTC on Friday, 13 November. Microsoft says the cause was a technical issue introduced during a network capacity upgrade, unrelated to any cyberattack or AI agent.

A Network Capacity Upgrade Caused the Microsoft 365 Outage

Microsoft attributed the outage to a technical issue introduced during a network capacity upgrade, BBC News reported on 13 November. Earlier in the incident, TechCrunch security editor Zack Whittaker reported, Microsoft’s status updates described a large number of servers becoming unexpectedly unavailable, and engineers at first could not find what was making servers in its data centers fail. Microsoft later said it had found no evidence of malicious activity and that the incident was not related to cyber activity or AI agents. By Friday the company said it had addressed the underlying issue and was restoring service to the remaining customers and regions. Microsoft told the BBC it will publish a full post-incident report once its review is complete.

The Outage Spread From Outlook and Teams to Copilot, Defender and Intune

The disruption started as an Outlook email and Teams problem and widened into consumer and business email, files and documents, Copilot, Defender security screening and the Intune admin dashboards, TechCrunch reported. Some users also could not reach their cloud-based Windows PCs. Microsoft first acknowledged that users in the Americas could not access Exchange Online mailboxes at 17:54 GMT on Thursday and soon warned the problem might affect Microsoft 365 globally, according to BBC reporter Rachel Clun. The BBC counted more than 70 affected countries. Affected institutions included the Japanese government’s shared IT platform, the Philippines Supreme Court, an Australian intelligence agency, and GP surgeries, hospitals, schools and police forces in the UK. OpenAI and Anthropic both reported problems linked to the outage. Microsoft’s website says more than 430 million people use the affected apps, though the number of people actually affected has not been confirmed.

Microsoft Restored Service in Stages So Servers Would Stay Up

Microsoft said it restored services gradually because of the size of the problem and to keep them stable once they came back. At 04:36 GMT on Friday it reported only partial recovery in some regions. By 13:32 it described broader improvements in North America, sustained improvements across much of Europe and early recovery signs in Asia Pacific. Doug Madory, director of internet analysis at the network monitoring company Kentik, said traffic began returning to Microsoft’s services after about 18 hours of downtime, TechCrunch reported. James Ferryman, a computer science professor at the University of Reading, told the BBC that the servers handling secure connection setup were being swamped, and that each failure pushed work onto the next server while huge numbers of devices kept trying to reconnect. “You can’t just turn it all back on at once,” he said.

Microsoft Extended the Email Timeout Window to 48 Hours to Protect Delayed Messages

During recovery Microsoft extended its email timeout window to 48 hours so that delayed messages would not expire before delivery caught up, according to BBC technology reporter Tom Gerken. Messages that arrived late show the time of delivery as their timestamp. Microsoft’s final status update says services across the affected environments fully recovered and the email backlog was processed. The previous major incident came in September, when TechCrunch reported that a multi-day outage of Microsoft’s cloud services was blamed on a core authentication configuration issue.

Why This Matters for Engineering Leadership

The fault behind this outage was a capacity change, and Microsoft said on Friday that it had addressed it. Much of the remaining time went to recovery: servers had to be brought back in stages while devices retried, and queued mail needed a longer expiry window to survive the wait. That makes the recovery path a design decision, and it deserves the same review as the change itself. For systems your team runs, a capacity upgrade should carry a rollback plan, and the restore plan should include client retry backoff with jitter, admission control on connection setup, and a staged regional restart that has been rehearsed. Queues should keep messages longer than your worst observed recovery time. For services you buy, this incident took down Intune and Defender alongside email, so the tools administrators would use to respond were also unavailable. Keep an incident communication channel and an admin runbook that work when your primary productivity vendor is down, and ask that vendor for its post-incident report before your next renewal review.

FAQ

What caused the Microsoft 365 outage in November 2026?

Microsoft says a technical issue was introduced during a network capacity upgrade. It says the incident was not caused by a cyberattack or by AI agents.

How long did the Microsoft 365 outage last?

Microsoft’s service health page lists a start time of 10:05 UTC on 12 November 2026 and an end time of 16:10 UTC on 13 November, a span of about 30 hours.

Which Microsoft services were affected?

Outlook and Exchange Online email, Teams, files and documents, Copilot, Defender and the Intune admin dashboards were all reported as affected. Some users also had trouble reaching cloud-based Windows PCs.

Were emails lost during the Microsoft outage?

Microsoft extended its email timeout window to 48 hours so delayed messages would not expire, and its final update says the email backlog was processed. Delayed messages carry their delivery time as the timestamp.

Why did recovery take so long?

Microsoft says it restored services gradually because of the scale of the problem and to keep them stable once they were back. A University of Reading computer scientist told the BBC that reconnecting devices can overload servers if everything returns at once.

Will Microsoft publish a post-incident report?

Yes. Microsoft told the BBC it will release a full post-incident report once it completes its review.

Source: Microsoft’s latest email outage turned into a full-blown global cloud crisis by Zack Whittaker, TechCrunch, November 13, 2026; Microsoft ‘deeply sorry’ for global outage as it reports broad recovery by Rachel Clun, BBC News, November 13, 2026; and Why is Microsoft’s outage lasting so long and will emails be lost? by Tom Gerken, BBC News, November 13, 2026. Primary source: Microsoft service health status, final update for the Microsoft 365 incident.

Read more news & analysis