Microsoft 365 Outage in South Africa: What's Down (13 Nov 2026)

Listen to this article

Microsoft 365 outage in South Africa, with an office laptop inbox failing to load and red alert lighting in front of the Johannesburg skyline at night

By Tihan van der Walt | 13 November 2026

Can't get into Outlook this morning? Don't bother resetting your password. And no, it's not Telkom or Vodacom either. Go and look at Microsoft's own service status page: it still shows a service degradation for Microsoft 365, and warns that users may be unable to connect to multiple Microsoft 365 services. My last look was at about 09:30 South African time on Friday 13 November 2026. Azure, on the same page, had no issues.

This has been going since last night. Microsoft's status account first flagged Exchange Online mailbox problems at 19:54 SAST on Thursday, initially in the Americas. Then, thirteen minutes later, it said the issue may affect Microsoft 365 connectivity for users in any region. That's the one that reached us.

Downdetector South Africa picked up a first spike in complaints around 20:00 on Thursday, then a sharp surge after 05:00 today as people logged on for work. Microsoft 365 hit 874 reports at 07:53. Outlook hit 744 at 07:38. Teams, 79 at 07:53. Copilot got a smaller spike. All before 08:00. I know, those are user reports and not a headcount of the mailboxes that are down, but they match Microsoft's timeline and what I've been hearing all morning.

What's actually broken

The useful split, and Microsoft has been making it too, is between anything that runs on Exchange Online and everything else. Exchange Online, for the non-IT folk, is the cloud system behind your business email, calendars, contacts and tasks in Outlook.

Going by a Microsoft update at 07:46 SAST, which The Citizen's LNN reported:

Outlook and email. Users may be unable to load inboxes, or may see delays, in the Outlook website, the desktop app and on mobile. Basically everywhere you'd open mail.
Teams. The good news: joining calls, video meetings, group chats and ordinary channel messages have been working. Anything leaning on Outlook, though, hasn't. That's scheduling meetings, sending calendar invites, viewing calendars, creating virtual events, and creating or deleting private or shared channels.
Copilot. Copilot Chat may not load or may not respond. It also can't pull information from work email, calendars or files.
SharePoint and OneDrive. Files sitting directly in SharePoint Online have remained accessible, but don't count on the “recently used files” list.
Admin and security tools. Spare a thought for your IT person, who may struggle to load the Microsoft 365 admin centre, the Microsoft Defender portals or Microsoft Purview. Safe Links has also been failing. That's what scans links inside your email, and when it fails it can stop you opening links you were sent.

There's no time from Microsoft for full restoration. Its latest public status post, at 08:12 SAST, said it was still working to recover frontend servers and improve capacity. And when it does come back, it's likely to be patchy. Microsoft has already said as much: as capacity comes back, users may see a gradual recovery as their clients reconnect. Some of you will be back before your colleagues. You're not imagining it.

How the night unfolded

I went back through Microsoft's official status posts, in South African time, and ignored the “the whole internet is down” headlines doing the rounds last night.

There was a separate, earlier Teams incident in North America, logged as TM1550488, opened at 17:40 SAST and marked fully resolved at 19:09. It is not this outage.

EX1550543, the Exchange incident, opened at 19:54. By 20:07 it was global. Then at 20:35 came “elevated load across multiple regions” on a portion of Microsoft's infrastructure, and a broader incident, MO1550821, was opened. If you can't get into the admin centre (and plenty of administrators can't), you can follow the public version of that on Microsoft's service health page.

The important clarification came at 21:31. In Microsoft's words, the elevated load was “not driven by an increase in traffic, but by a number of servers being unexpectedly unavailable,” which then strained the servers still standing. Shortly after that came word that some servers were failing automated health checks. It hadn't, it said, identified any recent change that correlated with the start of the impact. But the failed health checks may have come from a component that processes server health.

The rest of the night was fixes that didn't stick. At 23:00 Microsoft said an alternative health-check method had brought a subset of servers back. At 00:09 on Friday it admitted the actions so far had “not provided the anticipated relief.” A core process on affected servers was becoming overwhelmed. At 02:06 it said servers returned to service could become unresponsive again once they received traffic. At 05:13 it confirmed later changes were “not taking as expected,” and that it was looking at other ways to increase capacity in the face of “sustained traffic management imbalances.” At 06:36 it reported partial recovery in the regions where it had adjusted traffic protocols, redistributed traffic and applied “strategic traffic blocking measures.” And as of 08:12, it was still busy with frontend servers.

Microsoft has said plainly that it is “not aware of any malicious activity that is causing this incident.” It will publish a post-incident report once the problem is fully resolved. Till that report lands, I'd stick to what they've actually posted and no further.

This is not the whole internet

All sorts of other services lit up on Downdetector last night. Read some of the coverage and you'd think AWS, Google and Cloudflare had fallen over too. I'd be careful with that.

Amazon Web Services posted on X that its services were “operating normally,” and that an “unrelated disruption elsewhere on the internet is prompting some inaccurate news reports.” They pointed people at the AWS Health Dashboard, and CNBC and a few others picked the statement up. And an Amazon spokesperson told Macworld that AWS had been functioning normally all day.

Microsoft's own status page, when I looked this morning, had no Azure issues on it. The problem Microsoft is officially owning up to is Microsoft 365, and particularly whatever sits on Exchange Online.

Why does this matter? On a morning like this bad info travels faster than email. Today, that's not hard. So when a client sends you a post saying “AWS is down too”, check the operator's own status page before you forward it to anyone.

ChatGPT and Claude are up. Their Outlook connections are not.

If your ChatGPT or Claude tools that read mail, calendars or files have failed this morning, that does not mean those products are down.

OpenAI's official status page, at the time I checked it, was reporting elevated errors affecting the Microsoft 365 connector: new connections may be broken and tools are not working as expected. The rest of ChatGPT was not the subject of that incident. Anthropic's status page said the same kind of thing in more detail: mail, calendar and search tools in its Microsoft 365 connector were still failing for most requests, and Claude for Outlook remained affected, because of the Microsoft 365 incident.

In other words, the AI assistants themselves have been available. The plugs into Microsoft have not.

What I'd do this morning

Start with the obvious. This isn't a local IT problem, so don't treat it like one. Resetting your password in a loop won't fix it, and neither will an hour of reinstalling Outlook. Check Microsoft's status page or the @MSFT365Status account, tell your team once, and move on.

Then use what still works. Teams chat and calls have been the exception, and WhatsApp, SMS or a phone call will do what email can't until the queues clear. If a client is waiting on a proposal or a login link from you, I wouldn't wait for Outlook before telling them it's delayed.

Files are simpler. If SharePoint or OneDrive still opens via a direct link or a synced folder, work from there. Don't rely on “recent files.”

Be careful with links in email while Safe Links is failing. But don't assume every blocked link is an attack either. The scanning tool itself is part of what's broken.

Now the bit for marketers. If you run ads or send newsletters, this outage is about Microsoft 365, not Google Ads or Meta, and your campaigns are not automatically off. What may be off is your ability to approve copy, get a calendar invite or reply to a client from Outlook. Plan the work around that, not around a mythical “the internet is down” pause.

When service starts coming back, give it a few minutes. Microsoft has already warned that recovery will be gradual as clients reconnect. So if your inbox is back and a colleague's isn't, that fits what they've described. It doesn't mean your colleague has a unique fault.

The part that will still matter next week

For a lot of South African businesses, email is still the quiet glue. Invoices. Job offers. OTP messages, meeting links, the “please find attached.” Nobody notices until it stops. This outage is a Microsoft problem, not a local connectivity one. That's why a fibre line and a backup LTE SIM don't help if the mailbox lives in Exchange Online.

You don't need a second Microsoft tenant. You do need a way to reach staff and clients that doesn't depend on the same inbox. A shared WhatsApp group, a cellphone list, even a simple status note on your website. It sounds unfashionable until a morning like this.

In due course Microsoft will publish that post-incident report. I'll read it. Till then, what matters for a Joburg workday is above, current as of 09:30 SAST on Friday 13 November 2026.

Microsoft 365 is down this morning: what’s broken for South African businesses, and what still works