The story still heating up tonight is Microsoft. Not its earnings—the outage affecting Outlook, Teams, and Copilot, which began Thursday afternoon Eastern time and has now lasted into another day without being fully fixed.
When things first blew up on Thursday, reports about Outlook in the US alone on Downdetector briefly reached around 130,000, with one tally putting the number at 138,000. Within an hour, the outage had spread to 74 countries and territories. UK police, the NHS, schools, and Parliament were affected, as were Japan’s government office systems, the Singapore Government Technology Agency’s email, and the Supreme Court of the Philippines. Some airlines rerouted customer-service emails, and some companies postponed earnings calls. Microsoft assigned two incident numbers: Exchange Online was EX1550543, and the broader Microsoft 365 outage was MO1550821. Nearly every service tied to Exchange was hit.
It wasn’t a hack. Microsoft said it had found no signs of an attack, and Charles Lamanna, executive vice president of its Copilot business, said the same. Later, the company said a batch of servers had suddenly gone offline, leaving the rest unable to cope. A process responsible for establishing secure connections was overwhelmed, reducing capacity. They restarted it, but it went down again when traffic poured back in. At one point, they also tried rolling back a network configuration change. A patch was applied at the authentication and routing layers, but initially it was only “partially successful.” Some data centers in Asia-Pacific are now shifting traffic back, and some users can access their mailboxes again. Backlogged emails are also being delivered, but the service is still intermittent. The patch is being rolled out to other regions. The email timeout window was extended to 48 hours to prevent the system from treating backlogged messages as expired and discarding them. Microsoft still hasn’t given a timeline for a full recovery.
The stock market was less forgiving. Microsoft closed Thursday at $523, then fell to around $508.51 in premarket trading, down about 2.8%. Nasdaq futures barely moved over the same period, so this drop was Microsoft’s own. The company sells Copilot as an assistant that can check your unread messages and manage your calendar. But now the email system itself is down, and Copilot can’t connect either. Integrations connecting Anthropic and OpenAI to Microsoft 365 are also reporting errors, and Claude isn’t working in Outlook. A company selling smart office tools, with people still refreshing the login page the next day.
Office work, government services, and trading desks are all concentrated on the same cloud. One process responsible for establishing secure connections couldn’t cope, and communications for police, hospitals, and trading all went down at once. Reports say traders switched to phone calls. A recovery in Asia-Pacific doesn’t mean things will be working for the European and US morning sessions. Until the backlog is cleared, people will still be unable to open their email today. When US markets open tonight, whether the stock can hold around $508 will depend on how much longer the outage drags on.
#微软宕机 #Microsoft
When things first blew up on Thursday, reports about Outlook in the US alone on Downdetector briefly reached around 130,000, with one tally putting the number at 138,000. Within an hour, the outage had spread to 74 countries and territories. UK police, the NHS, schools, and Parliament were affected, as were Japan’s government office systems, the Singapore Government Technology Agency’s email, and the Supreme Court of the Philippines. Some airlines rerouted customer-service emails, and some companies postponed earnings calls. Microsoft assigned two incident numbers: Exchange Online was EX1550543, and the broader Microsoft 365 outage was MO1550821. Nearly every service tied to Exchange was hit.
It wasn’t a hack. Microsoft said it had found no signs of an attack, and Charles Lamanna, executive vice president of its Copilot business, said the same. Later, the company said a batch of servers had suddenly gone offline, leaving the rest unable to cope. A process responsible for establishing secure connections was overwhelmed, reducing capacity. They restarted it, but it went down again when traffic poured back in. At one point, they also tried rolling back a network configuration change. A patch was applied at the authentication and routing layers, but initially it was only “partially successful.” Some data centers in Asia-Pacific are now shifting traffic back, and some users can access their mailboxes again. Backlogged emails are also being delivered, but the service is still intermittent. The patch is being rolled out to other regions. The email timeout window was extended to 48 hours to prevent the system from treating backlogged messages as expired and discarding them. Microsoft still hasn’t given a timeline for a full recovery.
The stock market was less forgiving. Microsoft closed Thursday at $523, then fell to around $508.51 in premarket trading, down about 2.8%. Nasdaq futures barely moved over the same period, so this drop was Microsoft’s own. The company sells Copilot as an assistant that can check your unread messages and manage your calendar. But now the email system itself is down, and Copilot can’t connect either. Integrations connecting Anthropic and OpenAI to Microsoft 365 are also reporting errors, and Claude isn’t working in Outlook. A company selling smart office tools, with people still refreshing the login page the next day.
Office work, government services, and trading desks are all concentrated on the same cloud. One process responsible for establishing secure connections couldn’t cope, and communications for police, hospitals, and trading all went down at once. Reports say traders switched to phone calls. A recovery in Asia-Pacific doesn’t mean things will be working for the European and US morning sessions. Until the backlog is cleared, people will still be unable to open their email today. When US markets open tonight, whether the stock can hold around $508 will depend on how much longer the outage drags on.
#微软宕机 #Microsoft
