On Thursday 12 November, shortly before 6pm UK time, Microsoft 365 began failing for users around the world. Outlook was hit hardest, with Teams, SharePoint, OneDrive and Copilot also affected. More than 12 hours later, on Friday morning, Microsoft said service capacity had still not returned to normal.
What the reports say
Reports of Outlook problems peaked at nearly 140,000 around 18:00 GMT on Thursday, according to Downdetector, as reported by the BBC.
Microsoft said a portion of its infrastructure was "experiencing elevated load across multiple regions" and that servers had become "unexpectedly unavailable", straining the rest (The Verge, BBC).
Microsoft said it was not aware of any malicious activity. This was an internal failure, not a cyber attack (CNBC).
At around 9:15am London time on Friday, Microsoft said users may still be unable to use Outlook and any services that interact with email, with no timeline for a full fix (CNBC).
Microsoft reported improvements after applying a patch to an authentication and routing component, and described the outage as affecting millions of users for more than 12 hours (The Guardian).
A note on the charts circulating on social media: some show spikes for AWS, Google and Cloudflare alongside Microsoft. Amazon said AWS was "operating normally" and that reports of outages on its platforms were inaccurate (CNBC). This was a Microsoft outage, not the internet going down.
Why it matters for a small business
Nobody did anything wrong in thousands of affected offices. Their staff turned up, their laptops worked, and their own internet connections were fine. They lost a working day because one supplier's systems failed and there was nothing they could do but wait.
That is the real lesson. It is not that Microsoft is unreliable; large cloud services are usually very dependable. It is that when a business relies on a single outside provider for something essential, that provider's bad day becomes the business's bad day.
Our portability promise
Every application Lansafe AI Studio builds is designed to be portable. The customer owns it, and it can run on our hosting, on another provider, or on the customer's own hardware. It does not need any one cloud supplier to exist.
Put plainly: an application of ours running on a business's own server would not have been taken offline by Thursday's Microsoft failure, because it does not run on Microsoft's systems. Staff on the local network could have kept using it.
We also want to be honest about the limits:
Where an application deliberately connects to Microsoft 365, for example to send emails through Outlook, that one step would have been delayed. A well-built app queues the work and sends it once the service is back, instead of losing it.
Running on your own hardware brings its own responsibilities: power, backups, updates and security. Portability gives you the choice; it does not remove the need to look after the system.
No system is immune to every failure. The aim is to make sure no single outside supplier can stop your business.
Our first live example is LanGuard, which was developed on Lansafe hardware and now runs on a customer's own equipment. Read the LanGuard story or see how portability works.
Questions worth asking today
Which of our everyday tasks stopped yesterday evening, and which supplier did each one depend on?
If that supplier failed again for a full day, what would we lose?
Do we own the software and data behind our most important processes, and could we move them if we needed to?
If those answers are uncomfortable, talk to us. We will start by listening to how your business actually works.