---
title: "Microsoft outage took 74 countries offline for 12+ hours; failed health checks, not hackers"
description: "Servers failed automated health checks, then failed again under real traffic, stalling recovery across Outlook, Teams and Copilot"
author: "Luis Goa"
published: 2026-11-13T15:26:53Z
modified: 2026-11-13T17:28:48Z
url: https://rews.cc/a/microsoft-s-health-checks-killed-the-patient-inside-the-nove-1da240
language: en
tags: ["microsoft", "outage", "copilot", "infrastructure", "ai", "tech"]
publisher: "Rews (https://rews.cc)"
---

# Microsoft outage took 74 countries offline for 12+ hours; failed health checks, not hackers

*Servers failed automated health checks, then failed again under real traffic, stalling recovery across Outlook, Teams and Copilot*

By Luis Goa · November 13, 2026 · https://rews.cc/a/microsoft-s-health-checks-killed-the-patient-inside-the-nove-1da240

## In brief

- A Microsoft outage beginning 1:26 p.m. ET Thursday disrupted Outlook, Teams, Copilot and Microsoft 365 in 74 countries for over 12 hours
- Microsoft blamed servers failing automated health checks, not traffic or a cyberattack; restarted servers repeatedly failed again under load
- Downdetector reports for Outlook peaked near 130,000; such figures count user reports, not affected accounts
- Governments, courts, banks and airlines were hit; OpenAI’s and Anthropic’s Microsoft 365 connectors also failed
- Australia’s home affairs minister said Microsoft saw parallels to the 2024 CrowdStrike incident, but no root cause is confirmed

More than 12 hours after Microsoft’s outage began on Thursday, the company said it had achieved only *marginal success* restoring the services, [Al Jazeera reported](https://aljazeera.com/news/2026/11/13/microsoft-hit-by-global-outages-what-happened-and-why). By then, Downdetector had logged user reports from 74 countries and territories. The proximate cause, per Microsoft’s own status updates: servers failing automated health checks and dropping out of rotation. Not elevated traffic, and not, the company says, an attack.

[As Ashley Belanger wrote in *Ars Technica*](https://arstechnica.com/tech-policy/2026/11/microsoft-blames-failed-server-health-checks-not-hackers-for-global-outage/), Microsoft’s first explanation pointed at *elevated load across multiple regions*. The company later revised that: the load was real, but the cause was servers going unexpectedly unavailable, straining what remained. Four hours into the fix, Microsoft was still stuck: *We’re opening all available pathways to restore service, as our actions taken thus far have not provided the anticipated relief.* Its diagnosis sharpened to a *core process on affected servers* becoming overwhelmed.

## How the failure compounded

The recovery kept producing new failures. Servers Microsoft brought back online would, in the company’s words, *become unresponsive again once they receive traffic* — the classic signature of a restarting fleet that falls over the moment live load returns, forcing a slower cycle of isolating infrastructure, restarting components and reintroducing traffic in slices. With no quick fix coming, Microsoft extended its email timeout to 48 hours, an admission that delivery queues would backlog for up to two days.

The countermeasures, [as Security Affairs catalogued them](https://securityaffairs.com/202932/security/microsoft-cloud-outage-hits-outlook-teams-copilot-and-ai-services.html), were standard thundering-herd management: pause all deployments, redistribute connection traffic across worldwide infrastructure, shift traffic away from affected systems, and restart the dead servers to rebuild capacity. Microsoft’s admin-center notice MO1550821 listed Exchange Online, Teams, SharePoint Online, Microsoft 365 Copilot and Defender XDR as affected. The same notice history put the outage’s start at 1:26 p.m. ET Thursday — and it followed a separate, resolved Teams incident (TM1550488) that morning, 9:45 to 10:53 a.m. ET, which Microsoft traced to *resource contention* and which by its own account was unrelated.

![Three institutional buildings connect by cream cables to an unplugged connector beside a large envelope on a blue background.](https://rews.cc/img/21a39b63e56636401360a301742fb9865f5937f5.webp)

## Reading the numbers with care

The size estimates need their caveats attached. [The *Boston Globe* reported](https://bostonglobe.com/2026/11/12/business/microsoft-outlook-outage/) nearly 130,000 Downdetector reports for Outlook just after 1 p.m. ET; Security Affairs logged more than 97,000 at 2:07 p.m., plus 38,000 for Microsoft 365 and about 4,500 for Teams; [Euronews counted](https://euronews.com/2026/11/12/microsoft-outage-hits-outlook-teams-and-copilot-disrupting-thousands) 120,000-plus Outlook reports around 1:30 p.m. and over 46,000 for Microsoft 365. These are snapshots of user-submitted reports at different hours, not a census of affected accounts — one locked-out worker can file several reports, and one office of five hundred can file none.

What the geography shows better than the counts: [The Next Web’s timeline](https://thenextweb.com/news/microsoft-outage-outlook-teams-12-hours-global) has the failure starting in the Americas and spreading globally within an hour, with a tenfold jump in UK reports at 17:43 UTC, then surges in India and New Zealand as their workdays began. That propagation pattern — near-instant, worldwide, spanning an hour — points at shared control-plane infrastructure rather than any single region’s servers.

## What it wasn’t, and what it might be

Microsoft moved early to rule out the scariest hypothesis. Charles Lamanna, executive vice president for Copilot, Agents and Platform, posted on X:

![A navy gear containing a cream world map has a broken tooth with red fracture edges against a yellow background.](https://rews.cc/img/a01147eb7114e38ffb38deb610d5c41834ba71e7.webp)

> We regret that today’s outages caused disruptions for our customers, and we are working to address the issue. We will continue to share more information as soon as we have it, and for now I can confirm that there is no indication this is the result of malicious or cyber activity.

The likelier template is self-inflicted. Australia’s home affairs minister, Tony Burke, told reporters in Sydney that Microsoft believed it was dealing with something resembling the 2024 CrowdStrike incident — an update gone wrong — though Microsoft has not confirmed any root cause, saying only that it is examining affected servers and recent configuration changes. The comparison is worth keeping at arm’s length: the July 2024 CrowdStrike update bricked [about 8.5 million Windows machines](https://www.cnbc.com/2024/07/20/microsoft-says-about-8point5-million-of-its-devices-affected-by-crowdstrike-related-outage.html) — under one percent of all Windows devices, per Microsoft’s own estimate at the time — by crashing endpoints directly. This outage left devices untouched and took down the service side, which is a different failure class with different remediation: no fleet of machines to visit, but also no customer who can fix anything by rebooting.

![A dark blue briefcase with a yellow clasp breaks through a cream safety net against a teal background.](https://rews.cc/img/407557a0fd2e9b1a5ee9725cd22d64366e703ccc.webp)

## The blast radius

The downstream damage map reads like a dependency audit of Microsoft 365. Japan’s government reverted to paper processes after ministries couldn’t send email for much of Friday. Australia’s spy agency and federal police were hit. The Philippine Supreme Court warned of delays; Miami Beach and Austin told residents to use the phone; a Colorado sheriff’s office said bond payments for prisoners were delayed. Bankers at Morgan Stanley and Goldman Sachs in Hong Kong told CNN they lost email and Teams for the day. IndiGo and Akasa lost email support channels, and EV maker Workhorse postponed its earnings call. Standard Chartered and ING Australia warned customers that scammers were already impersonating tech support.

The failure propagated one hop further, into companies that aren’t Microsoft. OpenAI reported elevated errors on its Microsoft 365 connector — the pipe that lets ChatGPT read a user’s mail, calendar and SharePoint files — warning that new connections could fail. Anthropic saw Claude’s mail, calendar and search requests fail for most users, and traced it to an upstream provider before confirming the Microsoft incident. Both companies’ models stayed up; what broke was the integration layer. That is the structural fact of the episode: a single vendor’s health-check bug degraded products at two competitors, because the connectors that make AI assistants useful at work all terminate in Exchange and SharePoint.

By Friday, Microsoft reported *broader improvements in North America, sustained improvements across much of Europe, and encouraging recovery signals in Asia Pacific*, with intermittent issues and email backlogs still being drained. Downdetector reports had dropped substantially from Thursday’s peak. What has not arrived, a day in, is a root-cause statement — and the timing adds its own data point, since the outage ran while EU leaders met in Dublin to discuss reducing dependence on foreign technology providers. The number to watch next is not a report count but whatever Microsoft’s post-incident review says about why a health-check failure could take 74 countries of email down for half a day, and why restarted servers kept dying under the traffic they were restored to serve.

## Sources

- [Microsoft's global outage may be worse than Amazon's or CrowdStrike's, expert says](https://arstechnica.com/tech-policy/2026/11/microsoft-blames-failed-server-health-checks-not-hackers-for-global-outage/) — Ars Technica
- [Microsoft outage disrupts Outlook, Teams and Copilot in 74 countries](https://aljazeera.com/news/2026/11/13/microsoft-hit-by-global-outages-what-happened-and-why) — aljazeera.com
- [Widespread outage hits Microsoft apps](https://bostonglobe.com/2026/11/12/business/microsoft-outlook-outage/) — bostonglobe.com
- [Microsoft’s Outlook and Teams outage passes 12 hours](https://thenextweb.com/news/microsoft-outage-outlook-teams-12-hours-global) — thenextweb.com
- [Microsoft Cloud Outage Hits Outlook, Teams, Copilot and AI Services](https://securityaffairs.com/202932/security/microsoft-cloud-outage-hits-outlook-teams-copilot-and-ai-services.html) — securityaffairs.com
- [Microsoft outage hits Outlook, Teams and Copilot, disrupting thousands](https://euronews.com/2026/11/12/microsoft-outage-hits-outlook-teams-and-copilot-disrupting-thousands) — euronews.com