Last updated on
Is Your Business Ready for the Next Microsoft 365 Outage?
Microsoft 365 has had nine major outages since 2023. Here are four practical steps to keep your business running when the next one hits.
Last updated: 15 November 2026. Resolved. Microsoft restored service and closed the incident early on 14 November. Its full Post-Incident Report is due within five business days, and I’ll add the root cause here when it’s published.
If your inbox stopped loading on Thursday evening, you weren’t alone. On 12 November 2026, Microsoft 365 went down worldwide, and it took more than a day to fully recover. Outlook, Teams, SharePoint, Copilot and Defender were all affected (Windows Forum).
Microsoft is tracking the incident as MO1550821, which started at 18:26 GMT. It said a large number of servers had become unexpectedly unavailable, which pushed extra load onto the servers that were left. At the peak, Downdetector logged nearly 130,000 Outlook reports (Life News Agency). Microsoft paused deployments, moved traffic and restarted servers (ABC7). That didn’t work: restored servers went unresponsive again as soon as traffic reached them (Cybersecurity News). Things only began to stabilise after a patch to an upstream authentication and routing component, rolled out region by region (Windows Forum).
Resolved, 14 November: Microsoft’s final record puts the impact at 10:05 UTC on 12 November to 16:10 UTC on 13 November, about 30 hours, although most users felt it from Thursday evening. It closed the incident at 04:15 UTC on 14 November, once the email backlog had cleared. Its preliminary cause is “an issue during an earlier network capacity upgrade that surfaced following a recent network change”. Delayed email was held, not lost, and a full Post-Incident Report is due within five business days (Enrow). Microsoft says it was a technical issue, not a cyberattack (Forbes).
If you’re an admin, two things to do now:
- Check for missing mail. Run a message trace in the Exchange admin center for anything users say never arrived (Windows Forum).
- Claim your SLA credit. Credits aren’t applied automatically. Open a service request in the Microsoft 365 admin center quoting MO1550821 and EX1550543, with your impact times and number of affected users. The deadline is 31 December 2026, and if you buy through a partner, claim through them (LazyAdmin).
During the outage, Microsoft’s admin notice listed what still worked and what didn’t. Worth knowing for next time:
- Email is the core of it. Mailbox access through Outlook on the web, Outlook desktop and mobile may fail, and delivery may be delayed. Features in other apps that rely on Exchange Online are hit too.
- Teams mostly works. Calls, meetings, group chats and channel messages are largely unaffected. Calendar, scheduling and meeting invites are not.
- SharePoint files still open. Accessing files stored in SharePoint remains functional, but external sharing may fail.
- Protected content may not open. Purview services, including sensitivity labels and DLP, are impacted, so labelled and encrypted files may not open.
- Intune device enrolment, check-in and compliance policies have been intermittently impacted.
- Power Apps mobile users should stay signed in. Users who are already signed in can avoid impact by not signing out, clearing app data or reinstalling.
You can’t stop Microsoft having a bad day. What you can control is how much that bad day costs your business. This post covers the four things I’d sort out before the next one.
This wasn’t a one-off
In the last five years, Microsoft 365 has had at least nine major multi-service outages. Since January 2023, that’s one every five to six months. At least seven lasted four hours or more, and several took out most of a working day.
| Date | What went down | How long | Cause | Source |
|---|---|---|---|---|
| 12 Nov 2026 | Exchange Online, Outlook, Teams calendar, Copilot, Purview, Defender | About 30 hours (Microsoft’s final record) | Issue from an earlier network capacity upgrade, surfaced by a later network change (preliminary) | Windows Forum |
| 23 Jul 2026 | SharePoint, OneDrive, Teams, Power Automate, admin center | About 5 hours | Bug in automated network maintenance removed IP routes | BleepingComputer |
| 22 Jan 2026 | Outlook, Exchange mail flow, Defender, Purview | Almost 10 hours | North American infrastructure not processing traffic | The Register |
| 29 Oct 2025 | Microsoft 365, Outlook, Teams, admin center, Azure | 8+ hours | Accidental Azure Front Door configuration change | Petri |
| 1 Mar 2025 | Outlook, Exchange Online | Several hours | Faulty code change, rolled back | Medha Cloud |
| 25 Nov 2024 | Exchange Online, Teams calendar, mail flow | Most of a business day | Faulty change to token infrastructure | Medha Cloud |
| 30 Jul 2024 | Microsoft 365 apps, Azure | 8 to 10 hours | DDoS attack, made worse by a defence error | Medha Cloud |
| 5 Jun 2023 | Outlook, OneDrive, Azure portal | On and off for several days | DDoS attack | Medha Cloud |
| 25 Jan 2023 | Teams, Exchange, Outlook, SharePoint, OneDrive | About 5 hours | Bad WAN router command | The Register |
That count doesn’t include shorter regional incidents. It also leaves out the CrowdStrike update of July 2024, which took down 8.5 million Windows machines without being a Microsoft cloud fault. The lesson is the same for all of them: plan for a working day without Microsoft 365, roughly twice a year.
1. A backup way to talk to each other
When Outlook and Teams go down together, most businesses lose every internal channel at once. This time Teams calls and chats mostly kept working, but you can’t count on that next time.
Agree a backup channel before you need it, and write it into your disaster recovery and incident communication plan:
- A second business messaging platform, such as Slack or Google Chat, set up for at least your key people. Consumer apps like WhatsApp are tempting, but they sit outside your security, retention and compliance controls. Check that the backup doesn’t sign in through Entra ID, or it could go down with Microsoft 365.
- One named person, plus a deputy, who posts updates: what’s down, what to do in the meantime, and when the next update is due.
- A way to reach staff and customers that isn’t Outlook: a banner on your website, or your intranet if it’s still reachable.
If your phone system runs on Teams Phone, plan for calls going down too. The skeleton contact list in section 3 covers that.
2. Offline mode for the files that matter
A cloud outage doesn’t wipe your files. It only stops you reaching them, which is why a bit of local copying goes a long way.
- OneDrive “Always keep on this device”: set this on the handful of folders people can’t work without, such as price lists, templates and current client files. Those copies stay readable when SharePoint isn’t. Admins can push this setting to every device with Intune.
- Outlook in cached mode: the classic desktop app keeps a local copy of your mailbox, so staff can still search and read existing mail during an outage. Outlook on the web and the new Outlook can’t do this to the same extent.
- Desktop Office apps: Word and Excel open local files without a connection. Make sure people know where their synced copies live before they need them.
Get InfoSec on board for anything sensitive. An offline copy lives on a laptop, not in the cloud, so InfoSec should know which documents are cached and approve it, and those devices should be encrypted and managed. Bear in mind too that during this outage, files protected with sensitivity labels may not open even when they’re stored locally.
One more check: do you actually know which five or six files would stop the business if nobody could open them for a day? Write that list down. It’s the start of your continuity plan.
3. An emergency contact list that doesn’t live in Outlook
Your contacts, your supplier emails and your IT provider’s details are probably all in Outlook. That’s exactly where you can’t get at them during an outage.
Keep a short skeleton contact list somewhere else: printed, in a password manager, or as a PDF on key people’s phones. It isn’t a staff directory. It only covers the people who make decisions or fix things during an incident:
- Your key decision-makers: CEO, CFO, IT director, and other directors and executives
- Your IT support or managed service provider, with an out-of-hours number
- Your Microsoft partner and any support contract or tenant ID they’ll ask for
- Your domain registrar and DNS host, in case you need to change mail or web records
- Your key customers and suppliers, so you can tell them you’re having trouble with email
Review it every six months. A list with a leaver’s number on it is almost as bad as no list at all.
4. A way into the admin center when sign-in is struggling
Some outages hit authentication itself. If your admins sign in through single sign-on (SSO), a federation provider or a strict Conditional Access policy, they can find themselves locked out of the very tools they need to diagnose the problem.
This is what emergency access accounts (often called break-glass accounts) are for. Microsoft recommends having at least two. Each one should be:
- Cloud-only, using your tenant’s
.onmicrosoft.comdomain rather than a synced or federated account - Excluded from at least one Conditional Access policy that could otherwise lock everyone out, such as a location or device-compliance rule
- Protected with phishing-resistant MFA, such as a FIDO2 security key kept somewhere safe, since admin portals now require MFA anyway
- Monitored: set an alert for any sign-in, because these accounts should almost never be used
Be realistic about what they can’t do. If Microsoft’s own sign-in service is down, no account will get you in. That’s when you need to know where to check status without logging in:
- status.cloud.microsoft, Microsoft’s public status page
- @MSFT365Status on X
- The Microsoft 365 Admin app on your phone, if you’re already signed in
Your 15-minute readiness check
These are suggestions and best practices from my own experience, not a one-size-fits-all standard. Agree them with your stakeholders, InfoSec included, before you roll anything out.
- We have a backup business messaging channel that doesn’t depend on Microsoft 365 sign-in
- One named person, plus a deputy, posts updates during an outage
- Critical folders are set to “Always keep on this device”, with InfoSec approval for sensitive files
- Outlook desktop is running in cached mode
- A skeleton contact list of key decision-makers exists outside Outlook and was reviewed this year
- We have two emergency access accounts, tested in the last six months
- Everyone knows where to check Microsoft’s status without logging in
If you ticked fewer than five, the next outage will cost you more than it needs to.
Disclaimer
Everything in this post is general best practice drawn from my own experience, not a prescription. Every organisation is different: its size, risk appetite, regulatory obligations and existing tooling all shape what “ready” looks like. Treat these recommendations the way ITIL treats its guidance: adopt and adapt what fits, scale it to your needs, and leave out what doesn’t. Decide on your approach with your own stakeholders, InfoSec and compliance included, before making any changes.