On 2–3 June 2026, Exchange Online incident EX1331830 left messages undelivered across North America, Asia-Pacific and Europe for more than an hour at a stretch. It followed a separate Office-and-Teams file-opening issue on 1 June and a Teams problem spike mid-month — a reminder that even the most reliable cloud platform on earth has bad days.
The reflex is to ask whether it was an attack. It almost never is — these are infrastructure and configuration events inside Microsoft’s estate, and there’s nothing a customer could have patched to prevent it. Blaming the platform is also the wrong takeaway, because the platform’s uptime is still far better than anything an SMB could run itself.
The right takeaway is continuity design. A second communication channel for when Teams or Exchange is down, a documented “mail is delayed” process so staff stop frantically resending, and SLAs you actually understand. You can’t prevent a Microsoft outage. You can decide in advance how your business keeps moving through one.