Microsoft 365 Outage Lasts Hours: 130,000 Reports, Copilot and AI Services Affected
Baca dalam 60 detik
- Microsoft 365, Outlook, dan Teams lumpuh lebih dari sembilan jam dengan hampir 130.000 laporan gangguan dari 74 negara.
- Microsoft memastikan tidak ada indikasi serangan siber, tetapi dugaan mengarah ke salah konfigurasi server yang memicu beban berlebih.
- Insiden ini memperlihatkan rapuhnya ketergantungan korporasi pada satu penyedia cloud dan mendesak perlunya rencana cadangan.

Microsoft 365 productivity services, including Outlook, Teams, Word, Excel, and the Copilot AI assistant, suffered a massive outage lasting more than nine hours on Thursday and Friday this week. Data from monitoring platform Down Detector recorded a surge of nearly 130,000 problem reports from 74 countries, making it one of Microsoft's largest incidents in recent years. The company stressed there was no evidence of outside involvement, but early recovery efforts failed to bring services back to normal.
The outage was first detected around 13.00 US Eastern time. Besides office applications, enterprise services such as SharePoint and OneDrive were also affected, as were connectors used by AI services from OpenAI and Anthropic. OpenAI confirmed it was investigating a problem with Microsoft connectors that caused new connections to drop and tools to malfunction. Anthropic reported an outage from an upstream provider affecting their Microsoft 365 connectors.
In its official statement, Microsoft suspected the outage originated from settings on servers responsible for maintaining service availability. The company tried various steps, from rolling back the latest update to limiting data traffic to the affected servers. However, when the fix was applied, the servers again became unavailable as soon as normal traffic flowed. As of early Friday Greenwich time, the Exchange service—the backbone of Outlook email—remained the worst affected, while Teams was relatively more contained.
"We regret the disruption and are working hard to resolve it. There is no indication that this was caused by malicious or cyber activity," said Charles Lamanna, Executive Vice President of Copilot at Microsoft, via his X account.
Mark Gregory, an associate professor at the School of Engineering, RMIT University, assessed that the incident touched the network layer within Microsoft's systems and bore similarities to the misconfiguration case that crippled the Optus network in 2023. According to him, when some servers fail, users are forced to share a smaller pool of servers. "If part of the pool is lost, all users will access the remaining servers, and that overloads them. The servers then fall over, and a cascading problem occurs," he explained. He added that companies need to strengthen testing environments before applying configuration changes on production networks, as well as prepare fallback options.
From a corporate standpoint, this outage underscores the risk of concentration on a single cloud service provider. When Microsoft 365 goes down, many organizations lose access to email, collaboration, and document storage simultaneously. Gregory noted the importance of alternatives: "Many people keep a Gmail account; if Microsoft is down, they switch to Gmail. If Gmail is down, they go back to Microsoft." However, for companies whose entire workflow is tied to one ecosystem, emergency switching is not always smooth. This is where a digital resilience strategy is needed, covering service redundancy and contingency procedures.
For Indonesia, this incident is a reminder that dependence on global cloud services is not without risk. Many offices, educational institutions, and government agencies in the country use Microsoft 365 as the backbone of daily operations. An outage lasting hours can halt communication flows, delay decision-making, and disrupt public services that rely on corporate email. Although there were no reports of permanent damage, the event shows that digital sovereignty is not just about technology adoption, but also readiness to face provider failures.
Microsoft stated that recovery would be gradual as clients reconnect. The company is also reviewing whether rolling back a number of recent configuration updates could speed up the process. Going forward, the question is not only how quickly Microsoft restores services, but how far customers—including in Indonesia—are willing to invest in backup schemes and platform diversification so that operations do not come to a complete halt when the main provider stumbles.



