Technology

Microsoft 365 outage runs into a second day as Exchange load spikes

68 views

Microsoft's email and collaboration services were still recovering on Friday, more than a day after an outage that knocked out Outlook, Teams and Copilot for users across dozens of countries.

How the failure spread

Downdetector recorded about 140,000 reports from 74 countries and territories within roughly an hour of the first complaints on Thursday. Microsoft said users were "experiencing issues accessing Exchange Online mailboxes through multiple connection methods in the Americas" and that part of its infrastructure was "experiencing elevated load" that was likely causing connectivity errors.

The company began redistributing connection traffic across its wider global footprint to spread the load. Messaging and online meetings kept working for most customers, while mail, calendar details and Copilot did not. Microsoft published updates under incident EX1550543 in its admin centre.

Sixteen hours to one service

After about 16 hours the fault was narrowed to Exchange, the mail platform behind Outlook. Microsoft said it was seeing positive results as capacity came back. It has not said what caused the load to spike, and there is no public indication of an attack.

Governments and airlines were among the organisations reporting disruption. Staff at agencies that run on Microsoft mail switched to personal accounts and phone calls; carriers warned of slower responses at service desks.

The cost of one supplier

Microsoft shares fell during the session as the incident stretched on, a reminder that the company is selling Copilot and other AI tools on top of the same platform that just failed.

The outage also revived an old argument about concentration. Email is still the system of record for contracts, invoices and approvals at most large organisations, and a mail outage stops work that chat tools cannot absorb. Regulators in Europe have been asking whether the bundling of productivity software with cloud infrastructure makes these failures more widespread than they need to be.

For IT teams the practical lesson is the same each time: identity, mail and file access fail together when they run on one platform, so the recovery plan has to assume all three are down at once.

Source: BBC News