This story is currently available in англійська only — a translation isn't ready yet.
Microsoft blames global outage on three separate faults collapsing at once
According to the company's preliminary report, a latent software bug that had sat unnoticed in the system since July, a gateway going offline and a 14 percent loss of server capacity triggered the cascading failure.
ℹ️ Озвучка браузером · студійний голос ШІ незабаром

Microsoft has attributed the cause of last week's outage, which affected millions of users worldwide, to three separate technical faults. In a preliminary report published on Monday, the company said the failure emerged from a software bug that had gone unnoticed in the system since July, one of its internal gateways going offline, and a reduction in server capacity. According to the company's report, each of the three problems was manageable when taken individually; occurring at the same time, they led to a cascading failure.
Charles Lamanna, Corporate Vice President of Microsoft Copilot, Agents and Platform, described the details in a statement on X. The company normally routes data flow between systems through two separate gateways. One of the gateways failing to work for Teams caused traffic to shift to the other gateway, creating excessive load there.
The second development that amplified the failure came in server capacity. Because of another problem carried over from the previous day and unrelated to the gateway failure, more than 14 percent of Microsoft's total server capacity went offline. Rising data traffic combined with reduced capacity triggered the software bug that had gone unnoticed for months; the servers became overloaded and unable to respond.
The outage began on Thursday when users in the US could not access their Outlook accounts. When the failure spilled over into the next day, millions of users worldwide could not reach their email accounts and Office 365 services. Release procedures in prisons were disrupted, hospitals had to keep records on paper, schools could not communicate with parents, and some police departments could not receive crime reports.
Microsoft announced on Saturday that services had been restored and apologized to users. The company stated that the outage was not caused by any cybersecurity breach. The reason the repair took so long was that pending requests flooded the servers all at once as systems came back online, locking the servers up again. Lamanna said the recovery mechanisms that normally run automatically were not designed for a non-response situation of this scale. The fact that the software bug dated back several months also ruled out standard remedies such as rollbacks, staged rollout and feature flags. Microsoft will additionally prepare a final report containing all the details of the incident.
Sources — how we know this
- PrimaryEkonomim (Dünya)