Were your plans for a productive day interrupted by an Outlook inbox that wouldn’t load? We’ll tell you what’s going on, so grab a coffee, because it’s still happening.
And while you’re here, consider setting up monitoring for your most important services, so your users don’t have to go through the same disappointment.
On November 12, 2026, many Microsoft 365 users couldn’t get to their email.
Outlook on the web and the desktop app couldn’t connect to mailboxes, and new mail arrived late. Anything else that talks to Exchange Online started failing with it: calendar and scheduling in Teams, Copilot Chat, parts of SharePoint, Defender and Purview, and even some pages of the Microsoft 365 admin center.
At 11:30 UTC on November 13, more than 19 hours after the first mailbox errors, Microsoft still listed the incident as ongoing.
Its latest update says mailbox access and delayed email are improving, recovery is done in parts of Asia Pacific, and it’s now moving on to other regions. Microsoft has also extended the email timeout window to 48 hours, so delayed messages should still arrive.
What happened at Microsoft (the short version)
Microsoft’s first incident went up at 16:17 UTC. Users couldn’t connect to their Exchange Online mailboxes through Outlook on the web, Outlook desktop, Exchange ActiveSync or REST. At 18:26 UTC, Microsoft opened a second, broader incident covering every Microsoft 365 service that depends on Exchange Online. It first saw the problem in the Americas, then said users in any region could be affected.
The explanation changed a few times as the night went on.
Microsoft first pointed to “elevated load across multiple regions.” Then it said the load wasn’t coming from extra traffic. A number of servers had become unavailable, and the ones still running were straining to cover for them.
Later it said that “a process on the affected servers responsible for establishing secure connections becomes overwhelmed,” and started reverting a recent network configuration change that lined up with the start of the problem.
Recovery came slowly. Servers that Microsoft restarted could stop responding again once traffic returned to them. At 07:29 UTC on Friday, Microsoft reported progress:
We’re seeing positive results for server stability after applying an upstream authentication/routing configuration component patch to a portion of our infrastructure.
Microsoft says it has no evidence of malicious activity, and it will publish a Post-Incident Report with the root cause once the incident is mitigated. Until then, the cause isn’t confirmed.
Azure stayed out of it. The Azure status page showed no active events on Friday morning, and Microsoft’s own list of affected services covers Microsoft 365 only.
The timeline
| Time (UTC) | What happened |
|---|---|
| Nov 12, 14:45 to 15:53 | A separate Teams incident delays chats and presence for users in the Americas. Microsoft hasn’t said whether it’s related |
| 16:17 | Microsoft opens an Exchange Online incident: users can’t connect to their mailboxes |
| 17:52 | Anthropic reports errors in its Microsoft 365 connector |
| 18:15 | Zendesk reports mail delivery issues with its Microsoft Exchange connector |
| 18:24 | Our Is Outlook down? check fails for the first time |
| 18:26 | Microsoft opens a broader incident for Microsoft 365 services that depend on Exchange Online |
| 19:42 | Twilio reports email delivery delays to inboxes hosted by Microsoft |
| Overnight | Microsoft says a process that sets up secure connections is overwhelmed and reverts a recent network change |
| Nov 13, 07:29 | Microsoft reports “positive results for server stability” after a patch on part of its infrastructure |
| 08:07 | Our Outlook check records its 33rd separate incident, the last one so far |
| 11:04 | Microsoft reports recovery in parts of Asia Pacific and moves on to other regions. The incident is still open |
Who said what, and when
Microsoft kept posting updates all night, on the Microsoft 365 Status account on X and on its public status page, and each update said when the next one would come.
One line in those updates matters for IT teams: “Admins may be unable to load some admin center pages.” The admin center is where Microsoft 365 admins usually read Service health. So during this outage, the public status page and the X account were the places to look.
Other companies started posting before Microsoft widened its incident. Anthropic reported errors in its Microsoft 365 connector at 17:52 UTC, and Zendesk reported mail delivery problems with its Microsoft Exchange connector at 18:15 UTC. Both came before Microsoft’s broader incident at 18:26. Twilio followed at 19:42 UTC with email delays to Microsoft-hosted inboxes.
If your support team answers tickets in Zendesk from an Exchange mailbox, this outage reached you twice: once in Outlook, and again in your help desk.
What we saw from UptimeRobot
Our Is Outlook down? page checks outlook.com from North America. The check failed for the first time at 18:24 UTC with a connection timeout. That was two hours after Microsoft’s first mailbox incident, and two minutes before Microsoft widened it to all of Microsoft 365.
After that, it went up and down all night.
Between 18:24 UTC on Thursday and 08:07 UTC on Friday, that one check recorded 33 separate incidents: 16 connection timeouts, 13 “503 Service Unavailable” responses and 4 “400 Bad Request” responses. Most lasted a few minutes before outlook.com answered again. Together they add up to about 3 hours and 14 minutes of failed checks. As of 11:30 UTC, it hasn’t failed again since 08:07.
That fits what Microsoft described. Servers came back, took traffic and stopped responding again. From the outside, it looked like more than 30 short outages in a row.
Our November 12 outage page for Microsoft 365 lists all three Microsoft incidents from that day. The knock-on incidents have their own pages too: Anthropic, Zendesk and Twilio.
Getting ready for the next one
Nobody outside Microsoft could fix this one. For everyone else, the first useful answer was whether the problem sat with them or with Microsoft, because that decides who gets paged and what users hear.
UptimeRobot’s Third party monitoring gives you that answer as an alert.
Microsoft 365 is in the provider list, and you choose the components you depend on, such as Exchange Online, Microsoft Teams or SharePoint Online. When their status changes, the alert comes through the email, Slack or SMS contacts you already use for your own monitors, and it doesn’t count as downtime on them.
The same goes for the tools that broke along with Microsoft this time. Zendesk, Twilio and Anthropic are on the list too, along with more than 6,000 other providers.
Get 50 monitors with 5-minute checks, free, no credit card required.
