Skip to content
feed: live
>_ 0dayNews
microsoft

M365 Outage Drops Teams, SharePoint, Admin Center

Microsoft 365 outage July 23 took Teams, SharePoint, and the M365 Admin Center offline, stranding security teams without their primary management console.

M365 Outage Drops Teams, SharePoint, Admin Center
Photo: Raysonho @ Open Grid Scheduler / Grid Engine / Wikimedia Commons · CC0
loop Loop · Published · 1 min read

Microsoft 365 experienced a broad service disruption on July 23, 2026, simultaneously knocking out Teams, SharePoint Online, Excel Online, and the Microsoft 365 Admin Center. BleepingComputer tracked the incident as it developed; users across all three services reported access failures through the afternoon and evening UTC.

Teams and SharePoint going down is the expected headline. The Admin Center is the piece worth noting separately.

The M365 Admin Center is where privilege controls live — conditional access policies, Entra ID user management, Defender configuration, and compliance tooling. When the Admin Center is unavailable, security administrators lose the ability to inspect or modify those controls in real time. An organization running an active incident investigation during this window was effectively working without its management plane.

This outage ran concurrently with a separate Exchange Online mailbox quarantine issue that began July 20 — two distinct Microsoft 365 disruptions at the same time. Microsoft has not connected the two incidents.

Microsoft attributed neither disruption to a security breach. Both are classified as platform-side engineering incidents.

Resolution status: Check Microsoft’s service health dashboard at admin.microsoft.com → Health → Service health for current status on both incidents. Incident timelines and post-incident reviews will appear there before any public statement.

The operational point this week made visible: cloud-hosted runbooks and IR playbooks become unavailable at the moment you need them if the cloud is the problem. A copy outside M365 — a network share on separate infrastructure, an internal wiki on a different platform, even printed binders — is the specific thing worth verifying exists. That is not a new observation. This week produced a clean proof of concept.

Found this useful? Share it.