Outlook went down today. We did not.
Microsoft's November 12 outage took Outlook and Teams offline for hours. Here is what our own monitor saw on our side, and why the outage never reached us.
A little after 1pm Eastern today, Outlook stopped working for a lot of people. Downdetector peaked near 130,000 reports, Teams and Copilot went with it, and Microsoft said a number of its servers had become unavailable and the rest were straining under the load. For many businesses that meant an afternoon of no email.
Our afternoon was quiet. That is not luck, and it is not because we are smarter than Microsoft. It is because of how our services are laid out, and that part is worth explaining.
What our monitor saw
We run our own uptime monitor, and it keeps every check it makes. Between 1pm and 5:45pm Eastern it ran 714 checks across the 14 services it watches (this site, the helpdesk, client sign-in, billing, onboarding, notifications and the rest). All 714 passed. Average response time was 181 milliseconds. The same window yesterday averaged 179.
Email to our support address kept arriving through the afternoon and turned into tickets the way it always does. Nothing to fix, nothing to post on the status page.
Why it did not reach us
We do not run on Microsoft 365. That is the whole first answer, and it is less clever than it sounds. When one company runs your email, your chat, your documents and your sign-in, one bad afternoon at that company takes all of them at once. Today was that afternoon for a lot of offices.
Most of what we run lives at the edge. Our public services run on a global edge network rather than in one data center. A request is answered close to whoever made it, and if one location has a problem, traffic goes somewhere else. There is no single building where everything lives. We wrote about the early version of this in the stack behind SentinelGrid.
The parts that do need a machine have somewhere to go. A few things need real servers, so we run our own, and the important ones have a standby. If one of those machines stops answering, its work moves to another one. We keep moving more of it to the edge so less depends on any single box.
We watch it ourselves. Our monitor does not rely on anyone else's status page to tell us something is wrong. That is why we built it in the first place: a vendor's status page tends to go green last.
What we are not claiming
We are not saying we can never go down. Everything goes down eventually, including us, and anyone who promises otherwise is selling something. Our edge provider can have a bad day too. What we can say is that today's outage was a Microsoft problem, our services do not depend on Microsoft, and the numbers above are what actually happened.
We are also not saying Microsoft 365 is a bad choice. It is a good product, and plenty of businesses should stay on it. The lesson is narrower: know what depends on what.
The useful question for your business
If Outlook was down for you today, ask one thing: what else stopped because of it? If the answer is "email, chat, files, the shared calendar and the login to half our tools," then one provider is a single point of failure for your whole office. That is fixable, and it does not require leaving Microsoft. It might mean a second way to reach your team, a phone tree that lives outside the email system, or making sure the tools you cannot work without do not all sign in through the same door.
That kind of mapping is part of what we do for clients. If you want help working out what your business actually depends on, and what happens when each piece goes down, take a look at our managed IT.