If you couldn’t get into Outlook or Teams on Friday 13 November, you certainly weren’t alone.
It was a very Black Friday!
As we all know, Microsoft suffered one of the longest outages in its history, affecting Outlook, Teams, Copilot and other Microsoft 365 services for more than 24 hours in some places. Downdetector recorded more than 7.4 million reports worldwide.
Was it the biggest IT outage ever? That’s still unclear. The 2024 CrowdStrike incident still has a claim to that title, affecting around 8.5 million Windows computers worldwide.
It was interesting watching Microsoft trying to fix it.
Servers were restored, infrastructure rebooted and patches applied. Some recovered servers then failed again. At one point Microsoft described its progress as having only“marginal success.”
Microsoft has since said the problem originated during an earlier network capacity upgrade and surfaced following a later network change. They have been clear at stating that there is no indication it was a cyberattack. Nevertheless, we notice other major incidents. Around the same time, the US aviation system experienced serious technology problems, while Europe has recently seen cyberattacks involving internet-connected supermarket refrigeration. Email, aircraft and supermarket freezers seem pretty unrelated.
So what actually went wrong?
Microsoft has an enormous number of servers running around the world, and those servers constantly check whether they’re healthy. Think of it as a server asking itself: “Am I okay?”.
The current thinking is that something went wrong with that health-check process. Servers started deciding they weren’t healthy and took themselves offline, sending their work elsewhere.
Then came the domino effect. The servers still running had to pick up the extra work. Some became overloaded and dropped out too. Microsoft would bring servers back online, only for the same thing to happen again. Suddenly, a fairly routine technical problem had become a worldwide outage.
The fix was equally unglamorous. Microsoft had to take traffic away, fix the affected infrastructure, and then very slowly bring everything back online without starting the whole cycle again. That explains why recovery took so long. Microsoft’s official wording is that the problem was introduced during an earlier “network capacity upgrade” and surfaced following a later network change. It has ruled out a cyberattack or malicious activity as the cause.
FAQs
Was this the biggest IT outage ever?
No. It was enormous, but the 2024 CrowdStrike incident affected around 8.5 million Windows devices and disrupted airlines, hospitals, banks and other organisations around the world. Friday’s incident is better described as one of Microsoft’s most significant outages.
Was it a cyberattack?
Apparently not. Microsoft says there’s no indication malicious activity was responsible.
Should I still trust Microsoft 365, or consider another platform?
One very bad outage doesn’t suddenly make Microsoft 365 a bad platform. Moving to Google Workspace or another cloud provider doesn’t remove the possibility of an outage either. I certainly wouldn’t be dragging an old Exchange server out of retirement and putting it back in the office cupboard.
Will I get compensation from Microsoft for the outage?
Possibly. Microsoft’s service agreements provide for service credits where eligible services fall below their agreed availability levels. These are generally credits against the service, rather than compensation for lost productivity or other business costs.
We’re currently working through the process to determine the most effective way of identifying eligible clients and making the claims on their behalf. Microsoft says CSP credit requests are submitted through the provider channel.
Given the magnitude of the event, we expect Microsoft will aim to resolve this efficiently at scale. That is likely to take them a little bit of time to organise. We are starting to liaise with the distributors to get advice from Microsoft and determine how best to proceed. We’ll work through the eligibility and claims process and let affected clients know what happens next. and we will update this article as information comes to hand.
Therefore, if you get your Microsoft Business 365 licenses via Kinetics or OutSource, you shouldn’t need to do anything at this stage to claim for the Friday 13th outage.