The 60-second check, a 4-step test that separates a Microsoft outage from a local problem, the incident ID decoder, and a dated log of real 2026 incidents. Microsoft 365 has had two multi-hour global outages this year, January 22 and August 31, plus a five-day Outlook connection incident in September, so knowing where to look is not optional.
The fastest authoritative answer is status.cloud.microsoft, Microsoft's public status page, which needs no sign-in and updates during major incidents. Each source below answers a different question, so check them in this order.
| Where to look | What it tells you | Sign-in needed |
|---|---|---|
| status.cloud.microsoft | Major public incidents across Microsoft 365. If something big is broken, it shows here. | No |
| @MSFT365Status on X | Microsoft's fastest public acknowledgments, often with the incident ID before the status page updates. | No |
| Service health in the Microsoft 365 admin center (Health, then Service health) | Incidents affecting your specific tenant, including ones that never appear on the public page. The most complete source. | Yes, admin account |
| Downdetector | Crowd reports. Useful as an early smoke signal, but not authoritative; spikes happen for local ISP problems too. | No |
Tenant-scoped incidents are the trap for small businesses: the public page says all green while Service health shows an advisory that explains exactly what your team is feeling. If anyone in the company has the admin login, that dashboard is worth the 30 seconds.
Four checks, about 4 minutes, and you will know whether the problem is Microsoft, your network, or one machine. Run them in order and stop at the first step that works.
outlook.office.com, Teams at teams.microsoft.com, files at office.com. If the web version works and the desktop app fails, the service is fine and the problem is your app or device. Head to the fix library instead of waiting for Microsoft.Microsoft tags every incident with a two-letter service prefix plus a number, like TM1315837 for a Teams issue, and that prefix tells you the blast radius at a glance. These are the prefixes a small business actually sees.
| Prefix | Service | Real 2026 example |
|---|---|---|
TM | Microsoft Teams | TM1315837, the Teams location prompt loop on macOS, May 2026 (our fix page) |
EX | Exchange Online (email) | EX1469649, Outlook mobile and Mac unable to connect for five days, September 2026; EX1331830, mail delays across three continents, June 2026 |
MO | Microsoft 365 suite, multiple services at once | MO1465074, the August 31, 2026 authentication outage; MO1221364, the January 22, 2026 outage |
SP | SharePoint Online | Affects file libraries and intranet sites |
OD | OneDrive for Business | Affects file sync and sharing |
OP | Microsoft 365 apps (Word, Excel, Outlook desktop) | OP1470668, Excel copy and paste broken, September 2026 |
MP | Microsoft Purview (compliance, information protection) | MP1474363, protected documents failing to open in North America, September 18, 2026 |
CP | Microsoft 365 Copilot | CP1474143, pay-as-you-go breakdown unavailable in the admin center, September 2026 |
IT | Microsoft Intune | Device management, Company Portal, and enrollment |
An MO incident is the one that hurts: multiple services degrading together, often including the admin center itself, which is exactly when you most want to check Service health and may not be able to.
The honest answer is that during a confirmed Microsoft outage there is nothing on your machines to fix, and the costly mistakes are the things people do while waiting. Four rules.
If Microsoft says all clear and the 4-step test points at your side, the usual suspects are local, and they are the everyday work of IT support. The common ones: office internet or DNS trouble that only shows on your network, a lapsed payment or unassigned license blocking sign-in for one user, a conditional access or MFA lockout after a phone change, and a broken app on one machine, like the Microsoft apps on macOS Tahoe right now. Those are fixable today, no waiting on Microsoft.
A running log of significant incidents as they happen, newest first. Last log update September 23, 2026.
Since September 22 at 12:44 UTC, some users opening the account view in Outlook on the web get the authentication error AADSTS90002, "Tenant not found", caused by a code misconfiguration in a change meant to modernize the Outlook on the web header. Mail itself keeps working. Microsoft validated a fix and was preparing to deploy it on the evening of September 22. Source: Microsoft 365 service health incident EX1476802, September 22, 2026.
From September 18 to 22, Poly Microsoft Teams Rooms on Android devices in HP DirectConnect mode dropped calls after about 30 seconds with the message "Different Versions on the Touch Panel detected, ending call shortly." A Teams version update (1449/1.0.96.2026249711) left the touch console and the room device on mismatched versions. Microsoft pushed a configuration update and advises restarting affected devices. Resolved September 22 at 22:00 UTC. Source: incident TM1477123.
From September 9 at 04:30 UTC to September 14 at 04:00 UTC, some users could not connect from the Outlook mobile app or Outlook for Mac. Microsoft traced it to an update meant to improve mail flow that drove high CPU on mailbox database infrastructure, tried to revert it, hit blockers, and instead backported a fix that took five days to reach every server. Outlook on the web worked throughout and was the recommended workaround. Microsoft published a post-incident report on September 16. Source: incident EX1469649.
From September 8 at 10:00 UTC, a code regression in a service update broke copy, paste, drag, and duplicate in Excel 2024, 2021, 2019, and 2016, the perpetual-license versions many small offices still run. Microsoft shipped fixed builds version by version through September 22. Excel 2019 gets the fix automatically; Excel 2016 users must download the updated build from the Microsoft Download Center. Resolved September 22. Sources: incident OP1470668 and Microsoft's paste not working in Excel known-issue page.
Starting August 31, 2026 around 15:08 UTC, first tracked as Exchange incident EX1464935 and then widened to MO1465074, a misconfiguration stopped authentication components deploying across part of the Exchange Online infrastructure. Email delays and failures, sign-in problems, and broken search spread to Teams, OneDrive and SharePoint, Purview, Defender XDR, the admin center, Copilot, and Universal Print, with tens of thousands of Downdetector reports. Mailbox connectivity came back about 12 hours in; search and residual issues ran into September 2. Sources: BleepingComputer, TechCrunch, and the Microsoft 365 Status account, August 31 to September 2, 2026.
Since July 17, add-ins that admins made "Available" have been installing themselves for users across Outlook desktop, web, and mobile, because part of the Exchange Online add-in sync pipeline returns the wrong enabled state. Admins can clear an add-in with Set-App -Identity <AppId> -OrganizationApp -PrivateCatalog -DefaultStateForUser Disabled. Microsoft's permanent fix is being enabled for all users, with full resolution expected by October 5, 2026. Source: incident EX1441488, updated September 22, 2026.
Copilot chat and portal access failed for several hours on June 11, 2026. Microsoft restored service by rolling back a faulty deployment the same day. Reported by Windows-focused trade press, June 11, 2026.
Microsoft acknowledged the incident at 10:33 a.m. ET on June 2, 2026, with messages delayed more than an hour for some users, and expanded the affected scope the same evening. Microsoft advised admins against routing workarounds during recovery. Check EX1331830 in Service health for the closure summary. Reported by TechTimes, June 3, 2026.
A macOS security update stops the OS from saving the location permission choice for Teams, so the prompt returns endlessly on macOS 14, 15, and 26. Microsoft's fix deployed and its final incident update on June 25, 2026 confirmed the prompts had stopped. Source: NHSmail Support mirror of TM1315837, June 25, 2026. Workaround for any Mac that missed the fix on our Tahoe fix page.
Starting around 19:37 UTC on January 22, 2026, Microsoft 365 services degraded for roughly nine to ten hours into January 23, with about 15,000 Downdetector reports at the peak. Microsoft's post-incident report attributed it to elevated load on reduced capacity during North America infrastructure maintenance, with backup systems overwhelmed when traffic shifted. Reported by The Register, January 23, 2026.
Real Microsoft outages are dramatic but rare, a few significant ones a year. The daily work stoppage is the local stuff: the app that will not launch, the printer, the sync error, the one laptop that fell out of the support window. That is the work Anthony AI does autonomously, in the Microsoft Teams chat your company already uses, with Vision remote support when a fix needs hands on the machine: you paste a one-time code, watch every action live, get a full action log in the chat, and the session caps at 20 minutes.
Pricing: Core is $24.99 per seat per month with 200 chats and 20 Vision minutes ($0.99 per minute after). Pro is $49.99 per seat per month with 400 chats and 60 Vision minutes ($0.89 per minute after). Every new user starts with 25 chats and 10 Vision minutes free, no credit card. Details on the pricing page.
Microsoft 365 status page, status.cloud.microsoft, accessed September 23, 2026. · Microsoft 365 service health incident notices EX1476802, TM1477123, OP1470668, EX1469649, MP1474363, CP1474143, and EX1441488, as published to tenant admin centers, September 2026. · BleepingComputer: Massive Microsoft 365 outage causes auth issues, service failures (MO1465074), August 31 to September 2, 2026. · TechCrunch: Microsoft 365 outage drags on, but things are improving, September 1, 2026. · Microsoft Support: paste not working in Excel, September 2026. · Microsoft 365 Status account on X (@MSFT365Status), incident MO1221364 updates, January 2026. · The Register: Microsoft 365 outage drags on for nearly 10 hours, January 23, 2026. · Microsoft post-incident report on the January 22, 2026 outage (elevated load during North America infrastructure maintenance), January 2026. · TechTimes: Exchange Online outage EX1331830, June 3, 2026. · Microsoft 365 Admin Center incident TM1315837, May 2026. · Windows trade press coverage of the June 11, 2026 Copilot outage.
Start free: 25 chats and 10 Vision minutes, no credit card. No contract. Cancel anytime.
Get Anthony AIMore fixes in the fix library · On Windows? See every known issue in the September 2026 updates · On a Mac? See the macOS 27 Golden Gate guide or every known macOS Tahoe Microsoft app issue · See what IT support costs in 2026 · How Vision remote support works