A major Microsoft failure left tens of thousands of users unable to reach Outlook email, Microsoft Teams and the Copilot AI assistant on Thursday, with the company attributing the fault to servers that stopped passing their own automatic health checks.

Reports of Outlook problems on the outage tracker Downdetector climbed to about 130,000 shortly after 1 p.m. US Eastern time, Al Jazeera reported. Some users said mail was neither sending nor arriving; others could not open their inboxes at all.

The figure had fallen to roughly 77,000 by just before 3 p.m., and to about 31,000 by 8 p.m., according to CNN. CNN reports the disruption began shortly after noon and ran through much of the American working day.

What Microsoft said

Microsoft said on X that servers were unavailable and were failing automatic health checks, which put additional load on those still running. It said it was testing a configuration change to redistribute traffic and restarting the affected services.

“We’re working to safely return them to service,” the company said.

In a later update quoted by CNN, Microsoft traced the problem to “failed health checks” that may have originated “from a specific component on the affected servers that’s responsible for processing them,” and said it was restarting servers that “have become unresponsive.” It described itself as “aggressively pursuing all pathways to reduce impact.”

By 8 p.m. the company listed the affected products as those touching Exchange Online and email — Outlook, SharePoint, Teams, Microsoft 365 and Copilot.

How wide was it

The three filings do not describe the same geography, and the difference is worth stating plainly.

Microsoft’s own early notice, as quoted by CNN, referred to users “in the Americas,” and CNN’s headline treats the disruption as a US event. Al Jazeera reports Microsoft as saying the issues “could impact users in any region.” Hiru News described it as a worldwide breakdown affecting tens of thousands, attributing that to foreign media rather than its own reporting.

No filing reached for this report gives a country-by-country breakdown, and none mentions Sri Lanka specifically. Sri Lankan offices running Microsoft 365 would have been in the early hours of Friday morning local time during the worst of it.

Update: recovery under way, 74 countries affected, cyberattack played down

Microsoft is working to restore services and says a deliberate attack is unlikely, according to a Hiru News filing at 12.23 p.m. Colombo time on Friday.

The company said it was already working to return the affected services to normal. The disruption to Teams, Outlook and other services since Thursday was caused by a connectivity problem, and it has affected tens of thousands of users across 74 countries, Hiru reported, attributing the country figure to foreign media rather than its own reporting.

Microsoft also said the likelihood that the breakdown was a cyberattack is very low. The filing notes that more than 430 million people worldwide use Microsoft applications.

What this settles, and what it does not

The 74 countries figure is the first geographic scope this desk has been able to record, and it bears directly on the disagreement set out above. It is squarely at the wide end: Microsoft’s own early notice referred to users “in the Americas,” while Al Jazeera reported the company as saying any region could be affected. A 74-country footprint supports the broader characterisation.

Two cautions. The figure is attributed to foreign media, not to Microsoft, and Hiru does not name the outlet — so it is a second-hand number, not a company disclosure. And “tens of thousands of users” remains the same order-of-magnitude estimate reported on Thursday, almost certainly derived from outage-tracking services that count user reports rather than actual affected accounts. Against 430 million users of Microsoft applications, tens of thousands of reports is not a measure of how many people were actually unable to work.

The statement that a cyberattack is unlikely is the first time any filing reached for this report addresses the question. It still falls short of a root cause: “a connectivity problem” is consistent with the failed health checks described above, but no filing explains what caused them.

Still open: no filing gives a restoration time, confirms services are fully back, names the countries affected, or says whether any data was lost.

Update, Friday evening: ‘encouraging recovery progress’, email starting to flow

Microsoft said it was seeing “encouraging recovery progress” as users began reporting that email was arriving in their inboxes again, the Daily Mirror reported on Friday evening, Colombo time.

In a statement on X, the company reported “email delivery performance improving across a growing portion of affected Microsoft 365 infrastructure”, while warning that some users may still “experience intermittent issues”. (The Daily Mirror’s text renders the product name as “Microsoft 356”; the company’s service is Microsoft 365.)

Reports were mixed. Some social media users said they were receiving email again; others said they still could not access it. The filing cites CNN journalists among those reporting restored inboxes.

What this adds

This is the first statement in the sequence that points to measurable improvement rather than to work in progress, and the first to be framed around a specific service — email delivery — rather than the platform as a whole. It is also the first acknowledgement that recovery is partial and uneven: “a growing portion” of affected infrastructure, with intermittent issues expected to persist.

Two cautions carry over. The evidence of recovery is again anecdotal — social media reports and journalists’ inboxes, not a company metric. And the item carries no reporter byline and is attributed in the Daily Mirror’s metadata to “Administrator”, the marker this desk associates with submitted or syndicated material; the CNN references suggest it derives from CNN’s live coverage rather than from Colombo reporting.

Still open: no filing gives a full restoration time, confirms services are back to normal, names the countries affected, identifies a root cause beyond the failed health checks and “a connectivity problem”, or says whether any data was lost.

Update, Friday night: North America and Europe restored, Asia-Pacific improving

Microsoft says services have substantially recovered in North America and across much of Europe, and that conditions in the Asia-Pacific region are also positive, Hiru News reported at about 8.16pm Colombo time on Friday.

The filing restates the scope at at least 74 countries, where businesses, schools and government institutions had their daily operations badly disrupted, and names Outlook, Teams and Copilot among the services users could not reach. It again attributes the country figure to foreign media rather than to Microsoft.

A sharper cause statement

This filing goes further on cause than any earlier one in this sequence. Microsoft is reported as explaining that the servers used to establish secure internet connections received more requests than they could handle, and were therefore unable to process those requests properly.

That is the first account that identifies a function — secure connection establishment — rather than a symptom. It is consistent with, and more specific than, the “failed health checks” and “a connectivity problem” described earlier, though no filing yet explains what drove the surge in requests.

Microsoft again emphasised that the breakdown was not the result of a cyberattack. That is now the second filing to carry the denial, and it is stated more firmly here than in Friday lunchtime’s “very low likelihood” formulation.

Still open: no filing confirms full global restoration, gives a completion time, names the 74 countries, or says whether any data was lost. This desk has removed the breaking flag from this article as the incident moves into recovery.

Update, Saturday: Microsoft apologises, and a user figure appears for the first time

Microsoft has apologised for the outage, Hiru News reported at 6.38am Colombo time on Saturday. This is the first filing in the sequence to carry an apology rather than a status note.

It is also the first to put a number on the users affected: Hiru gives around 430 million users in 70 countries, attributed to foreign media rather than to Microsoft.

That figure answers a question this article has carried as open since Thursday. It should be read with two cautions. The country count has moved between filings — Hiru’s own earlier report said at least 74 countries, and this one says 70. Neither number is sourced to Microsoft, and no filing explains the difference. The user total has no stated basis at all.

Timeline and remaining work

By Friday night, close to 24 hours after the breakdown began, most services were returning to normal, Microsoft is reported as saying. The company said it could not give a definite time for full restoration of all services, and that work was continuing.

The services named again are Outlook and Teams, with email, online meetings and institutional communications the functions most disrupted for businesses, government bodies and private users.

Still open: no filing confirms full global restoration, explains what drove the surge in connection requests, reconciles the 70-versus-74 country figures, substantiates the 430 million user estimate, or says whether any data was lost.

Not reported

Microsoft did not say when it expected services to be fully restored, and no filing reports what caused the surge in connection requests, an independently verified total number of users affected worldwide, or whether any data was lost.