Incident sheet · /security
- System
- Security & Risk
- Filed
- Read
- 3 min · 647 words
- Severity
- Routine
Microsoft 365 outage: 80% of gateway servers down, security gaps under review
Microsoft traced last week's global Microsoft 365 outage to a July defect; at peak, over 80% of affected gateway servers were down. Security teams now audit the gaps.

Event log
- A software defect introduced in July contributed to the global Microsoft 365 outage, Microsoft said Monday.
- At peak, more than 80% of the affected gateway's servers were unresponsive; 14% of capacity was already offline.
- Defender and Purview functions depending on Exchange Online were disrupted.
- A final post-incident review is expected later this week.
- Security vendors urged audits of detection gaps, MFA exceptions and emergency access accounts.
More than 80% of the affected gateway's servers stopped responding at the worst point of last week's global Microsoft 365 outage, Microsoft confirmed Monday. A software defect introduced in July helped trigger the cascade, and the incident now has enterprise security teams auditing what their monitoring missed while the platform went dark.
Charles Lamanna, presenting Microsoft's preliminary findings, identified three contributing factors:
- a latent defect tied to lock contention;
- increased traffic after Teams switched networking gateways;
- an unrelated issue that had already taken nearly 14% of gateway server capacity offline.
Together, these caused requests to accumulate until servers stopped responding. Traffic redirected to surviving servers spread the failure. Safeguards intended to contain problems between infrastructure partitions failed.
What did recovery actually require?
Engineers had to restore a previously working code path, take groups of servers offline, apply the fix, restart them, and only then reintroduce traffic. Microsoft admitted its automated recovery processes were not designed for such widespread unresponsiveness. A final post-incident review is expected later this week.
The outage was not a security breach. That distinction matters little to CISOs now asking whether detection, access controls and incident response held up when the communications provider itself failed.
What should security teams check first?
"The first thing I'd check is what your detection stack didn't see," said Hayden Covington, associate director of security operations at Black Hills Information Security. If endpoint or email security tooling was degraded, there may be a window in which telemetry and alerting ran thin or disappeared entirely.
"That window should be reviewed after the fact," Covington said.
Microsoft's own notices described disruptions affecting features that depend on Exchange Online, including functions within Defender and Purview. So the blind spot is not hypothetical.
Darin Fredde, senior director and offensive security evangelist at Ridge Security, said responders should verify three things: that an endpoint can generate an alert, that the alert reaches the response team, and that the endpoint can be isolated. "A quiet dashboard means little if the systems feeding it were offline," he said.
For access controls, Fredde recommended reviewing emergency accounts, temporary administrator privileges, MFA exceptions, and changes to device compliance requirements.
Who will your employees believe?
Disrupted communications hand attackers an opening. "Social engineering is a particular concern here, especially vishing and impersonation of Microsoft or internal IT support teams," said Matt Hull, vice president of cyber intelligence and response at NCC Group.
Hull warned that attackers could use phone calls and text messages to request password resets, multifactor authentication approvals or remote access. Employees should verify unexpected contact through established channels before acting on it.
Security teams also need to find out where company data went while the usual tools were unavailable. "I'd want to know two things: what did people do to keep working, and what was missed while monitoring was degraded," Covington said. "Workarounds often mean personal email, personal file sharing, or relaxed policies."
That shadow IT, created in hours under pressure, can outlive the outage by months.
Can you respond when the platform is down?
A separate communications or security product offers limited protection if employees still need the unavailable platform to reach it. Fredde said organizations should run an exercise that makes selected Microsoft services unavailable, then test responder communications, recovery instructions, credentials, log access and containment capabilities.
Emergency Microsoft accounts can address some lockouts, he noted, but they still depend on Microsoft's identity service. Break-glass glass, in other words, stored inside the same building.
After restoration, the checklist is concrete: review delayed alerts, remove unnecessary permissions, and assign every remaining monitoring gap an accountable owner and a deadline.
"Organizations need a minimum response capability that remains usable when the Microsoft services they depend on are unavailable," Fredde said. The final incident review lands this week. The internal one, for most enterprises, has barely started.
via techtarget.com (Source)
More from Tom Whitfield
Show full bio
Market editor covering people, launches and day-to-day practice in enterprise information technology and the leaders who run at CIO Monitor.
185 articles
The newsroom digest
One email a day with the stories that matter. No noise.
Cross-references
- Microsoft 365 outage traced to July networking gateway bug
- Microsoft 365 outage traced to three-month-old gateway bug
- Microsoft's Global Outage: A $7 Billion Lesson in Cloud Concentration Risk
- Microsoft 365 outage reportedly affects 100,000 users worldwide
- Microsoft promises full review of global M365 outage