On Thursday afternoon, November 12, Microsoft 365 started failing. Outlook wouldn't load mail, Teams calls dropped, and SharePoint and OneDrive came and went. Microsoft logged the start of the incident at 1:26 PM Eastern. By midday Friday it was still describing service as improving "across a growing portion of affected infrastructure." Some people were still seeing delayed email and patchy access, and Microsoft hadn't said when everything would be back to normal.
Microsoft's explanation so far is that "a number of servers" became "unexpectedly unavailable," which strained the rest of its infrastructure. It says there's no sign of a cyberattack. Its engineers paused deployments, moved traffic around and restarted servers. They also extended how long Microsoft holds undelivered email to 48 hours, so delayed messages wait in line instead of bouncing back to the sender.
That's the news. Here's what's worth thinking about once your inbox catches up.
Size isn't a backup plan
Most small businesses choose Microsoft, Google or Amazon for good reasons. They're everywhere, they're well run and they almost always work. But "almost always" carries a lot of weight. Thursday wasn't a first. On October 29, 2025, what Microsoft called an inadvertent configuration change in Azure Front Door took down Azure services and, with them, parts of Microsoft 365, for hours.
When a provider that size fails, it fails for everyone at once. Your support ticket joins a queue with millions of others. The status page tells you what Microsoft knows, when Microsoft decides to say it. Nobody there knows that your payroll goes out by email on Fridays, or that your front desk can't check anyone in without Teams.
None of this means you should leave Microsoft 365. It's good software, and for most offices it's still the right choice. The mistake is treating a big name as the whole plan, as if the size of the company you buy from were a guarantee.
What a local IT partner changes
Nobody outside Microsoft could have fixed Thursday's outage. NubiGate couldn't either, and anyone who tells you otherwise is selling something. What a partner who knows your business can do is make an outage like this an annoyance instead of a shutdown:
- A copy that lives somewhere else. Your mail, files and key documents are backed up outside the service that just went down, so the important things stay within reach.
- A plan for "email is down." Who gets told, how customers reach you and how your team keeps working, all written down before you need it.
- A person who answers. When something breaks, you talk to someone who already knows your setup, not a ticket number.
- Eyes on your systems. Good monitoring means we can often tell you what's going on, and what to do about it, before your staff starts calling.
- Fewer single points of failure. Not every part of your business has to depend on one provider's sign-in.
Five questions to ask this week
If Thursday cost you a day, put it to use. Ask whoever runs your IT, even if that's you:
- If our email went down for a full day, how would customers reach us?
- Do we have a copy of our mail and files that doesn't depend on Microsoft being up?
- When did we last try restoring something from that copy?
- Who on our team knew what was happening on Thursday, and how did they find out?
- Which of our tools stop working if one account or one sign-in stops working?
If the honest answers are "they couldn't," "not sure" and "never," you're in good company. Most small businesses have never been asked. It's also very fixable.
Choose the partner, not the logo
Going with a big provider is fine. Choosing it only because it's big, and stopping there, is where the risk is. What matters is who looks out for your business when that provider has a bad day, and whether they had a plan before it happened.
That's the work we do at NubiGate: managed IT, backup and disaster recovery, help desk and security for businesses without an IT department of their own. If this week's outage hit you, talk to us. We'll show you where you're exposed and what it would take to make the next outage a non-event.
Sources: Microsoft's incident updates as reported by BleepingComputer, and Thurrott on the October 2025 Azure outage.
