Microsoft 365 had three separate incidents on 21 October 2026, hitting SharePoint Online migrations and Teams SIP calling. The longest, a migration failure, ran for 33h 52m before Microsoft confirmed a fix was fully deployed, and the fault sat entirely on Microsoft's side.
Started
10:10 UTC
Duration
Lasted 1d 9h
Source
IsDown
Next time
Microsoft 365 is degraded. Is your product working?
An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.
50 monitors for your own websites. No card.
Affected components: Microsoft Teams, SharePoint Online
What happened?
The first incident started at 10:10 UTC and affected anyone migrating through the Migration Manager Platform (MMP) API, whether from the Microsoft 365 Admin Center or a third-party tool. Migrations run through the older SharePoint Migration Tool (SPMT) were not affected, a distinction Microsoft repeated in every update. Microsoft's own updates show a long back-and-forth: by 12:48 UTC they said they'd identified a fix and were validating it, by 18:51 UTC that validation was still running and taking longer than expected, and by 22:49 UTC they admitted the fix itself had integration issues and had to be redone. A root cause was eventually named: a configuration issue had taken down the components that support MMP migrations. By 15:58 UTC the following day, deployment had reached roughly 75% of Microsoft's internal environment, with production still pending. The incident closed out at 33h 52m, matching the 8-hour-plus median this service tends to run for its issues over the past 90 days. Separately, two SIP calling incidents hit Teams: one from 19:13 UTC lasting 1h 19m, and a second from 20:15 UTC lasting 20h 39m, both involving SIP devices disconnecting on sync and being unable to log back in, while Teams web, desktop and Teams Phone Devices calling stayed unaffected. During the window, users in the UK, France and Denmark reported a related but distinct problem: Power Automate flows not loading, logged at 11:21, 10:34 and 10:12 UTC respectively. Microsoft's status page does not tie this directly to the migration incident, though the timing overlaps closely.
Learning
This was a failure inside Microsoft's own migration and calling infrastructure, not anything on your end, and no amount of checking your own endpoints would have surfaced a configuration fault in Microsoft's MMP API or SIP call routing. The user reports about Power Automate flows failing to load show the same pattern: a backend dependency broke somewhere in Microsoft's stack while your product kept functioning normally from the outside. Microsoft's own fix attempts stumbled twice, with a validated patch later found to have integration issues, which tells you these incidents can run long even once Microsoft says they've found the cause. If you depend on MMP migrations or SIP-based Teams calling, your monitoring of your own systems would show nothing wrong because nothing on your side was wrong. This is the gap external monitoring is built to catch: it watches the provider's side continuously, so when Microsoft's status page is still the only account of what broke, you have an independent read on when service actually comes back. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.