On 23 July (incident MO1437424) a networking update took Teams, SharePoint, OneDrive, Power Automate, Copilot Chat, Loop and the admin centre offline for roughly 11 hours, mainly in North America, before Microsoft reverted it. Then from 31 August (EX1464935, later MO1465074), what Microsoft described as core authentication configuration issues inside Exchange Online spread to Teams, Graph, Purview, OneDrive, SharePoint, the admin centre, Copilot, Universal Print and Defender XDR. It was fully mitigated on 3 September — about 72 hours after it began.
Microsoft’s reliability is good. It isn’t perfect, and when it fails it fails for everyone at once. The August event is the one to think about: an authentication problem doesn’t stay inside one service, because every Microsoft 365 workload trusts the same identity plumbing.
A plan B doesn’t need to be elaborate. Know where you’ll check service health. Have a way to reach staff and customers that doesn’t depend on Teams or Exchange. And keep an independent backup of mailboxes, OneDrive and SharePoint — not for outages, which Microsoft fixes, but because the next incident might be the one where data goes missing.