Where are we now? Microsoft 363? 362? We've lost count because Exchange Online isn't playing nicely this morning

Updated Microsoft's Exchange Online service fell over in the early hours of this morning.

The company's status orifice initially figured that the problem mainly affected users in India as its engineers noted the wobbling at around 0700 BST. Just under an hour later Microsoft had to admit it was another global outage.

It is the latest in what appears to be a battle of who can annoy their users more. Azure suffered a major outage earlier this week. Rival Apple then hit back with its own wobble before Microsoft continued the TITSUP* tit-for-tat this morning.

The mystery issue afflicted apps using Exchange Online protocols, including Outlook on the desktop, mobile devices, and "those dependant on REST functionality," Microsoft said. The company was taking a long hard look at what it might have changed in recent days that might have broken something.

Initial thoughts centred around a traffic routing problem, meaning that the gods of downtime would only bless those unfortunate enough to find themselves on the afflicted infrastructure.

Microsoft eventually pinned the blame on a "recent configuration update", rolled it back and, at time of writing, was "monitoring the service" for signs of life.

In the meantime, folk were left twiddling their thumbs and pondering the wisdom of a leap cloudwards versus an on-premises service, staffed by engineers that could be yelled at face to face.

Users reported problems sending and receiving mails, accessing folders and attachments, or even being able to log into their services. Some noted difficulty synchronising between Azure Active Directory and Exchange Online while there were also isolated reports of SharePoint and Teams struck by the curse of bork.

Not all users were affected, but for some it is the latest in a series of cock-ups at Microsoft's end, which will only serve to undermine faith in its cloudy productivity services. Email is, after all, a critical communication medium (no matter how much Slack wishes it wasn't).

While Microsoft has yet to explain itself (we suspect the postmortem for this week is going to be a big'un), users have turned the speculatogun on to the company.

Aha, so the DEV team is testing in production..

— Vishal Chotoe (@ViC_0508) October 1, 2020

Let they who have not suffered the bowel-loosening sensation of accidentally pushing dev code to production cast the first stone. Not that we'd dream of suggesting such slipshod antics from the Redmond gang.

Instead, we'll leave you with an emission from its marketing tentacle, mere hours before the wheels came off.

Exceed customer expectations—even during times of rapid change—using digital sales and service solutions. Learn more: https://t.co/bgBIdYF2Wd #ResilienceAtWork pic.twitter.com/Ul7gs8I5XS

— Microsoft 365 (@Microsoft365) September 30, 2020

Well, the company has certainly exceeded expectations in terms of outage count. Or has it?

A hat-tip to Register reader David for the heads-up.

Microsoft sent us a statement:

“Our engineers are addressing an issue that affected some Microsoft 365 services. Services are starting to improve as a result of the measures we have implemented.” ®

Updated at 1437 BST on 1 October to add

Microsoft has updated us further: "Our engineers have mitigated the issue for Microsoft SharePoint and Microsoft Teams users. The majority of users of our Exchange services are also seeing service recovery as we work towards full resolution."

Hallelujah!

* Total Inability To Serve User Protocols