An issue with load-balancing software caused messages to pile up, with some still undelivered on Tuesday morning
A day-long outage affecting Microsoft’s cloud email services appeared to have been resolved by late Monday, even as some customers continued to report continued problems or that messages affected by the outage hadn’t yet arrived as of Tuesday morning.
Outlook.com, the free web-based email service formerly known as Hotmail, was worst affected, with Exchange Online users also reporting problems on Monday morning. The Exchange issues were resolved as of around midday on Monday, according to Microsoft’s Office 365 status Twitter feed.
Exchange Online is a hosted business email service provided by Microsoft as part of its Office 365 cloud services suite.
“Intermittent connectivity is affecting customers in some European countries, which we are working to resolve as soon as possible,” Microsoft said in a statement.
The problems began just after 9 a.m. on Monday, according to Downtime Detector, which collates reports of service outages. More than one thousand users reported problems during the course of the day.
Most of the site’s reports originated from the UK and northern Europe, but reported issues extended across Europe to the eastern Mediterranean and beyond, with a few users from Jordan, Japan and Australia also registering problems.
Nearly half of the reports, 47 percent, said messages weren’t being received, while 37 percent reported login problems and 14 percent registered general website issues.
As of Tuesday morning a few users in the UK and Europe said they still hadn’t received messages from Monday.
In a message on the Office 365 status website dated 2:25 a.m. BST on Tuesday morning Microsoft said the problems with Outlook.com had been resolved, but acknowledged continued delays to delivering some messages.
“We’ve verified that new mails are being sent and received as expected, and the previous queue has returned to normal,” the company said. “A very small number of messages sent during the incident may be delivered over the next few hours as the last remaining messages are processed.”
In a message on Monday afternoon Microsoft said a “subset of infrastructure” wasn’t able to process requests as expected, causing service levels to drop, with the issue addressed by redirecting traffic to alternate infrastructure.
An update that evening said some load-balancing infrastructure had experienced “increased CPU utilization which does not appear to correlate with increases in user traffic”, which Microsoft addressed by making configuration changes. Following these measures Microsoft said the queue of undelivered messages began to drop over the course of of Monday night.
Twitter users expressed their vexation with the problems, with one user declaring, “I realise now that this is my punishment for still using #hotmail since 2000,” and another writing, “Yes, people do still use #hotmail in 2017.”
Microsoft says Outlook.com has more than 400 million active users.
Usage of cloud services has taken off amongst businesses as well as consumers, but the trend has meant increased disruption when the large remotely-based infrastructure services experience service issues.
In March Microsoft’s US East Coast cloud region experienced an outage that knocked a broad range of services offline, including Outlook.com, Office 365, Skype and Xbox Live.
About three weeks earlier websites including the US Securities and Exchange Commission, Adobe, Medium, Slack, Splitwise and Trello experienced difficulties after Amazon’s S3 cloud storage service suffered an outage.
The Amazon S3 problems also affected the US East Coast region.
How well do you know the cloud? Try our quiz!