TechPastWeek logo

Microsoft 365 Went Down for Hours After Servers Failed Their Own Health Checks

Outlook, Teams, SharePoint and Copilot went down for Microsoft 365 users worldwide on 12 November after servers dropped out of automated health checks.

An aerial view of Microsoft's Redmond campus in Washington state, where the company's cloud services are headquartered.
On this page

Key takeaways

  • Microsoft reported trouble accessing Exchange Online mailboxes in the Americas at 12:54 p.m. ET on 12 November, then widened the warning to users in any region 13 minutes later.
  • The outage also affected Teams, SharePoint Online, Microsoft 365 Copilot, Defender XDR, Planner, To Do, Intune and Power BI, according to Microsoft's status page.
  • Microsoft said a number of servers became unexpectedly unavailable and failed automated health checks, straining the remaining infrastructure. It found no evidence of an increase in traffic or a cyberattack.
  • Downdetector recorded nearly 130,000 reports of Outlook problems shortly after 1 p.m. ET. Microsoft has not published a user count.
  • The disruption came two days after Microsoft's November Patch Tuesday fixed 1,094 vulnerabilities, including an exploited Outlook spoofing flaw, but the company has not linked the two.

Microsoft 365 suffered a widespread outage on 12 November that knocked out business email, chat and collaboration tools for users in multiple regions, and the company's own explanation points to a failure inside its server fleet rather than a traffic surge or an attack.

Related reading: A Router Flood Halted Every U.S. Departure for More Than Four Hours · CoreWeave's Revenue Jumps 168% as Its Backlog Reaches $132.5 Billion.

The first public notice came at 12:54 p.m. ET, when Microsoft reported "issues accessing Exchange Online mailboxes through multiple connection methods in the Americas." Thirteen minutes later it widened the warning to users in any region. The company's internal incident record began earlier, at 16:17 UTC, according to reporting that compared the public posts with Microsoft's admin-center timeline.

By the afternoon, the list of affected services had grown to include Exchange Online, Microsoft Teams, SharePoint Online, Microsoft 365 Copilot, Defender XDR, Planner, To Do, Intune, Defender for Office 365, some Purview services and Power BI. Microsoft tracked the Exchange disruption under EX1550543 and the broader incident under MO1550821.

What Microsoft said

The company's explanation was technical and incomplete. In an update before 3 p.m. ET, Microsoft said affected servers were failing automated health checks and that it was testing a workaround that bypasses the component responsible for processing those checks. It was also investigating elevated processor usage on the affected machines.

"We determined the elevated load isn't driven by an increase in traffic, but by a number of servers being unexpectedly unavailable; straining the remaining infrastructure," Microsoft said in one status update. "We're working to safely return them to service."

That distinction matters. A traffic spike is an operational problem that can be absorbed with capacity. Servers dropping out of a health-check system can cascade: the remaining machines take more load, slow down, fail their own checks and drop out too. Microsoft said it began redistributing connection traffic across its broader global infrastructure to balance the load, and by about 4 p.m. ET the first affected servers were returning.

The status page still listed the incident as open at 21:41 UTC, which suggests a gradual recovery rather than one clean fix. Microsoft did not publish a root cause, a count of affected users or an explanation for why the health checks failed.

The user side

The outage was visible far beyond Microsoft's status page. Downdetector recorded nearly 130,000 reports of Outlook problems just after 1 p.m. ET, according to ABC7. Universities and businesses sent notices to staff. Some Google Workspace users also reported trouble reaching their email through Outlook desktop and mobile clients, a second-order effect of the Outlook failure rather than a Google outage.

Microsoft said it found no delivery failure for messages sent to Outlook users, but that is a narrow claim. The outage blocked people from opening their mailboxes, not necessarily from receiving mail. For an hour or more, a large share of the world's office workers could not get to email, chat or files through the tools they use all day.

The numbers are reports, not users

The 130,000 figure is Downdetector's count of problem reports, not a count of affected accounts. Microsoft has not said how many users or organizations were hit. The list of affected services also changed through the afternoon, so a service missing from one update may still have been degraded at another point.

Why a health-check failure cascades

The phrase "failed automated health checks" sounds minor. In a large cloud service it is not. Health checks are how load balancers decide which servers can accept traffic. When a server fails a check, it is taken out of rotation. That is the correct response to a sick machine. It becomes a problem when many machines fail at once, because the surviving servers have to absorb the traffic, and the extra load can push them past their own thresholds.

Microsoft's description fits that pattern. It said the elevated load was not caused by more users but by a number of servers becoming unavailable, which strained the rest. It also said it was investigating elevated processor usage on the affected servers. The company tested a workaround that bypassed the component processing the health checks, a sign that the checks themselves were part of the failure path rather than a neutral observer.

The recovery sequence was equally typical. Microsoft redistributed connection traffic across its global infrastructure, returned servers to service gradually and watched for re-failures. That is safer than switching everything back at once, but it means customers experience a long tail of partial recovery: some mailboxes work, some do not, and the status page says the incident is still open.

What administrators should take from it

For enterprise customers, the outage reinforces a few practical rules. The first is that cloud email needs a fallback plan even when the provider is Microsoft. That can be a secondary connection method, a documented process for using webmail or a phone-based incident channel that does not depend on the same identity provider.

The second is that status pages are not incident reports. Microsoft's updates gave customers a timeline and an incident number, but not a root cause or a user count. Administrators who need to explain the outage to their own leadership will have to wait for a post-incident review, if Microsoft publishes one.

The third is that patch cadence and operational stability are separate risks. A large Patch Tuesday does not cause an outage by itself, and there is no evidence that this one did. But the same engineering organization is managing both the update pipeline and the health-check system, and both are part of the reliability story. The 1,094 fixes this month are a measure of how much code is changing underneath a service that customers expect to be stable.

Microsoft's service-level agreements typically cover availability, not the quality of the explanation. Customers can claim credits for downtime, but credits do not recover the lost hour. The real lesson is the one cloud providers repeat after every major incident: redundancy is a chain, and the health-check layer is one of its links.

Timing and the Patch Tuesday question

The outage came two days after Microsoft's November Patch Tuesday, which fixed 1,094 vulnerabilities across its products. Those updates included CVE-2026-124194, an Outlook spoofing flaw that Microsoft said attackers had already exploited. The flaw could expose a user's NTLMv2 hash if the victim opened a specially crafted message file, allowing an attacker to authenticate as the user.

It is tempting to connect the two events. Microsoft's service advisories did not. The company said it was investigating the cause and had disclosed no evidence of malicious activity. SecurityWeek noted the timing and the absence of a stated link, which is the right way to read it: a coincidence worth watching, not a finding.

The Patch Tuesday context does add to the pressure on Microsoft's operations. The company fixed a record 1,413 vulnerabilities in October and more than 5,000 this year, according to SecurityWeek. Running a cloud service at Microsoft's scale means patching continuously while keeping millions of tenants online. The outage is a reminder that the health-check and failover layers are part of that operational surface, not an abstraction.

Why one outage matters

Microsoft 365 is one of the few software products with a genuinely global blast radius. A single incident can stop email at banks, hospitals, schools and governments at the same time. Most organizations have learned to treat cloud email as reliable enough to need no local fallback. Days like 12 November test that assumption.

The incident also shows how much of the recovery story depends on transparency. Microsoft gave customers a timeline, an incident number and a partial cause. It did not say how many users were affected, what specifically broke the health checks or whether the same failure could recur. Those are the questions enterprise administrators will ask before they decide whether to change their redundancy plans.

For now, the service is recovering. The more useful question is whether Microsoft will publish a post-incident review. The company has done so for some past outages, but not consistently. Without one, customers are left with a status page that says the incident is closed and no explanation of why their email disappeared for an afternoon.

What remains unknown

Microsoft has not published a root cause, a user count or a post-incident report. The 130,000 figure is Downdetector reports, not affected accounts. The reported start time of 16:17 UTC comes from a third-party comparison of Microsoft's status updates, not from a Microsoft statement in the coverage reviewed here. No link to the November Patch Tuesday has been established.

Tags

microsoftmicrosoft-365outageexchange-onlinecloud-infrastructure

Frequently asked questions

What caused the Microsoft 365 outage?
Microsoft said affected servers were failing automated health checks and dropping out of service, which left the remaining infrastructure to handle the load. The company also investigated elevated processor usage on those servers. It has not published a root cause.
Which Microsoft services were affected?
Exchange Online, Outlook, Microsoft Teams, SharePoint Online, Microsoft 365 Copilot, Defender XDR, Planner, To Do, Intune, Defender for Office 365 and Power BI were listed at various points during the afternoon.
Was the outage a cyberattack?
Microsoft has disclosed no evidence of a cyberattack. It said the elevated load was not caused by more traffic but by servers becoming unavailable, and it was investigating why the health checks failed.
Is Microsoft 365 back to normal?
Microsoft began returning servers to service around 4 p.m. ET and said it was redistributing connection traffic across its global infrastructure. Its status page still listed the incident as open at 21:41 UTC, so recovery was gradual rather than a single fix.
Share

Found this useful? Share it with your team.

TPW Editorial

TPW Editorial

Editorial Team

TechPastWeek's editorial team reports and edits every story we publish. We research each piece against primary and credible secondary sources, and a human editor reviews and signs off before it goes live.

Keep reading