Contents
On the afternoon of Thursday 12 November, Outlook stopped working for millions of people. Emails would not send, inboxes would not load, and Teams, SharePoint and Copilot went down with it. By Friday the 13th the outage was all over Instagram and X, mostly as jokes about an unexpected day off. Behind the memes is a more serious question: how does one of the world’s biggest email systems fail for the best part of a day, and what actually broke?
Now we know what broke. A week on, Microsoft says the cause was a software flaw it had added back in July, which sat unnoticed for more than three months until a busy Thursday afternoon set it off. At the worst point, more than 80% of the servers involved stopped responding.
Here’s what Microsoft says happened, why something that small could take down so much, and the twist in its own explanation.
The short answer
A hidden defect in a July fix for the servers that route Microsoft 365 traffic made them freeze under certain loads. On 12 November, Teams traffic was moved onto a system that also handles email while part of that system was already offline. The defect kicked in, the remaining servers were swamped, and safeguards failed to stop it spreading worldwide. It was not a cyberattack, Microsoft says.
Microsoft’s explanation: a bug that waited since July
Microsoft executive Charles Lamanna set out the cause in posts on X, Fox Business reported on 16 November. Three things combined:
- A hidden flaw. “A latent defect caused servers to become overwhelmed and unresponsive under certain load conditions,” Lamanna wrote. The flaw was in a fix written for Microsoft 365’s gateway servers, the machines that route traffic to its services. It had run worldwide for more than three months without problems, and Microsoft’s test labs didn’t reproduce the mix of traffic and hardware that later triggered it.
- More demand on one system. Teams traffic shifted onto a system that also supports Outlook email.
- Less capacity to absorb it. About 14% of that system was already offline because of a separate issue.
Then the safety systems made things worse. Microsoft’s automatic traffic redirection piled work onto the remaining healthy servers, and another problem stopped safeguards from containing the failure to one part of the network, so it spread globally. “At its lowest, more than 80% of servers were unresponsive,” Lamanna wrote.
“At its lowest, more than 80% of servers were unresponsive.””
Why recovery took so long
Servers that came back online were swamped by waiting traffic and failed again, which matches what Microsoft was saying in real time (see below). The faulty change also couldn’t be undone quickly with normal rollback tools. Engineers applied a fix and restarted servers in groups, returning them gradually. Teams, OneDrive, SharePoint and Office could switch away from the faulty infrastructure and recovered earlier. Exchange Online, which runs Outlook email for businesses, could not, so email was last back.
The twist
On the day, Microsoft insisted the extra load “isn’t driven by an increase in traffic”. Its later explanation does involve more traffic, though: Teams traffic moving onto the email system. Both can be true. Overall demand from users didn’t jump, but traffic was redirected onto one system that was already short of servers, and that’s where the flaw went off.
Lamanna said a final review would give more detail. Below is how the outage unfolded, in Microsoft’s words at the time.
What happened
Microsoft first said it was “investigating reports of users experiencing issues accessing Exchange Online” shortly before 1pm US Eastern time on Thursday, according to NBC News. That was shortly before 6pm in the UK. Exchange Online is the service that runs Outlook email for businesses and organisations.
It quickly spread beyond email. Microsoft’s advisory listed Exchange Online, Teams, SharePoint Online, Microsoft 365 Copilot and the Defender XDR security product, BleepingComputer reported. Microsoft tracked the problem under two incident numbers, EX1550543 for Exchange Online and MO1550821 for the wider Microsoft 365 suite, according to SecurityWeek.
At the peak, more than 120,000 people in the US reported Outlook problems to the tracking site Downdetector, according to ABC News in Australia. The National, using Downdetector data, reported that complaints came from 74 countries and territories within about an hour, and that reports across five Microsoft services reached 19,650 a minute, against fewer than 15 on an ordinary Thursday.
| Item | times the usual level |
|---|---|
| Outlook | more than 2,000× |
| Microsoft 365 | more than 100× |
| Teams | more than 100× |
| Copilot | more than 100× |
| OneDrive | more than 100× |
Downdetector counts user reports, not confirmed failures. Figures are minimums as reported.Source: The National, using Downdetector data. Chart: The Sidelight.
What Microsoft said as it happened
Microsoft’s explanation changed as engineers learned more. Taken together, its updates describe a chain of failures rather than a single fault.
Step one: servers dropped out. Microsoft first described “elevated load” on part of its infrastructure across several regions. It then corrected itself. “We determined the elevated load isn’t driven by an increase in traffic,” it said, but by “a number of servers being unexpectedly unavailable,” which was “straining the remaining infrastructure,” BleepingComputer reported. In other words, this was not a rush of users. Part of the system went missing, and the rest had to carry its work.
Step two: the remaining servers struggled. By mid-afternoon US time, Microsoft said affected servers were failing automated health checks and that it was investigating elevated processor usage on them, according to SecurityWeek. SecurityWeek also reported that a recent network configuration change had coincided with the start of the disruption, and that Microsoft had begun rolling it back. Microsoft did not say the change caused the outage.
Step three: a core process was overwhelmed. Around 4.40pm Eastern time, Microsoft said a “core process on affected servers is becoming overwhelmed, causing capacity loss,” NBC News reported. Later it was more specific: the process that establishes secure connections was becoming overwhelmed, so servers stopped handling requests, according to SecurityWeek and ABC News in Australia.
Step four: restarts did not hold. On Thursday evening Microsoft said it was returning servers to service and that they “remain stable so far.” It then warned that they “can become unresponsive again once they receive traffic,” ABC News reported. Just after 11.30pm Eastern time it said it was seeing “partial recovery across regions” where it had made adjustments, but not “sustained recovery.”
Step five: a patch that helped. In an update at about 2.30am Eastern time on Friday, Microsoft said it was “seeing positive results for server stability” after applying a patch to an “upstream authentication/routing configuration component” on part of its infrastructure, NBC News reported.
Microsoft said the load “isn’t driven by an increase in traffic” but by servers being unexpectedly unavailable.”
Why the fix took so long
Microsoft has not described the failure in these terms, but its updates match a pattern engineers call a cascading failure. When a group of servers drops out, their work moves to the servers that are left. If those servers are pushed past what they can handle, they slow down, fail their health checks and are taken out of service too, which pushes even more work onto the next group.
That also explains why simply restarting servers did not work. A server that comes back online is immediately hit with a backlog of connections from millions of email apps trying to reconnect. If the part of the server that sets up those connections cannot cope, it falls over again. Microsoft’s own description of restored servers becoming “unresponsive again once they receive traffic” is exactly that.
To break the cycle, Microsoft said it paused all deployments, moved traffic away from affected infrastructure, restarted servers and reintroduced traffic gradually, according to ABC News and The National. The Friday patch to an authentication and routing component, the part that decides how connections are checked and where they are sent, was the first change Microsoft described as giving positive results.
What Microsoft has not yet explained is the first link in the chain: why so many servers became unavailable at once. That is the question its promised report will need to answer.
Was it a cyberattack?
Microsoft says not. It said it was “not aware of any malicious activity that is causing this incident,” NBC News reported. In Australia, Home Affairs Minister Tony Burke said Microsoft had told him it did not believe a cyberattack was responsible, according to ABC News in Australia.
SecurityWeek noted that Microsoft had also just patched an Outlook security flaw in its monthly updates, but that Microsoft’s service advisories had not linked that flaw to the outage.
Who was affected
Microsoft has about 450 million users of its workplace software, NBC News reported, and the effects reached well beyond office inboxes.
- Governments. Japan’s government said its IT infrastructure was affected, and the US House of Representatives said its users were hit, according to ABC News in Australia. Australia activated a cyber agreement to prioritise its interests in the restoration.
- Public services. New York City Emergency Management warned that city services relying on Microsoft could be delayed, ABC News reported. In Colorado, the Arapahoe County Sheriff’s Office said bail had to be paid in person at the jail until Outlook was restored, according to NBC News.
- Businesses and schools. In Australia, shops lost email sales, telehealth appointments were disrupted and some distance-learning schools closed for the day. Even AI services were affected: OpenAI and Anthropic both reported problems with the tools that connect their assistants to Microsoft 365.
Why it took over Instagram
The timing helped. The outage ran into Friday the 13th, and by Friday morning in India the jokes were trending. The Tribune collected the typical posts on Instagram and X: requests to go home because Outlook was down, suggestions to switch to carrier pigeons, and the idea that Friday the 13th had turned out lucky for once.
The humour points at something real. For many workplaces, email, chat, files and meetings all now run through one company’s cloud. When it fails, there is often no other way to reach colleagues, which is why so many people found themselves with nothing to do.
Where things stand
Service has been restored. Microsoft has given its explanation of the cause, set out above, and says it was “not a cyberattack or an issue caused by AI tooling”, Fox Business reported. A final review is still to come, and we will update this article if it changes the picture.
Frequently asked questions
Why did Outlook go down on 12 November?
Microsoft says a hidden flaw in a July update to its Microsoft 365 gateway servers made them freeze under heavy load. It was triggered when Teams traffic moved onto a system that also runs email, while part of that system was already offline.
Was the Outlook outage a cyberattack?
No, Microsoft says. It found no sign of malicious activity, and Lamanna said it was not a cyberattack or caused by AI tools.
Why did Teams come back before Outlook?
Teams, OneDrive, SharePoint and Office could be switched away from the faulty infrastructure. Exchange Online, which runs Outlook email for organisations, could not, so it took longest to recover.
Updated 20 November 2026 with Microsoft’s explanation of the cause. The account of the outage as it happened was first published on 13 November.
References and sources
- 01Microsoft reveals hidden bug behind global Outlook and Microsoft 365 outage · Fox Business · 16 November 2026
- 02Microsoft services hit by major outage across the world · NBC News · 13 November 2026
- 03Massive Microsoft 365 outage disrupts Outlook, Teams, Copilot worldwide · BleepingComputer · 12 November 2026
- 04Microsoft 365 Outage Disrupts Outlook, Exchange Online and Teams · SecurityWeek · 12 November 2026
- 05Microsoft Outlook outage stretches on for hours · ABC News (US) · 13 November 2026
- 06Outlook email outage: Microsoft continuing to work on problem with ‘positive’ results seen · The National · 13 November 2026
- 07Microsoft outage continues, disrupting Outlook, Teams for businesses and governments · ABC News (Australia) · 13 November 2026
- 08Friday the 13th turns lucky? Microsoft, Outlook outage gives office workers unexpected holiday · The Tribune · 13 November 2026
House byline
Articles under this byline are researched and drafted with AI assistance, with every figure checked against its source before publication. The editor, Pankaj Guleria, reviews them and corrects them where needed.
Edited by Pankaj Guleria, Founder and Editor




