Microsoft has confirmed its Azure cloud services are back online after a major outage impacted services across multiple regions.
It's a tough time to be a cloud provider, with the Azure incident coming hot on the heels of an AWS outage that knocked a wide range of companies and services offline last week.
According to Microsoft, the Azure outage was caused by a CDN configuration error, lasted eight hours, and saw a wide range of Microsoft customers affected, including Alaska Airlines, websites for Heathrow, NatWest, and more.
Even Microsoft's own 365 suite was affected by the outage while proceedings at the Scottish Parliament in Edinburgh were paused due to technical difficulties.
In a service update, Microsoft said error rates and latency are “back to pre-incident levels”, but warned that a “smaller number of customers may still be seeing issues” akin to the turbulent recovery period experienced by AWS customers last week.
The outage was due to a configuration fault in a content delivery network (CDN) system. As with the AWS incident, once the outage started, the interlinked nature of these systems meant the disruption quickly cascaded into a wider problem.
What happened with the Azure outage?
Microsoft has detailed its initial take on the outage on the Azure status page, noting that the incident started at 3.45pm UTC and was spotted within 20 minutes, with manual recovery beginning at 6.45pm. The outage was resolved by midnight.
The outage affected customers and Microsoft services using Azure Front Door (AFD), which is Microsoft's content delivery network service, causing latency issues, timeouts and errors.
"An inadvertent tenant configuration change within Azure Front Door (AFD) triggered a widespread service disruption affecting both Microsoft services and customer applications dependent on AFD for global content delivery," the update said.
Microsoft added that this wasn't spotted because protection mechanisms failed due to a software bug.
"The change introduced an invalid or inconsistent configuration state that caused a significant number of AFD nodes to fail to load properly, leading to increased latencies, timeouts, and connection errors for downstream services."
As the nodes failed, they dropped out of the global pool, meaning traffic distribution across the nodes that remained became unbalanced. Microsoft noted that this exacerbated the outage and expanded the impact, meaning some regions that weren't affected by the AFD issues also saw service degradation.
To recover, Microsoft blocked all configuration changes and rolled back to the "last known good" version, which required reloading across a large number of loads. The recovery was slow to avoid overloading systems as they came back online.
"This deliberate, phased recovery was necessary to stabilize the system while restoring scale and ensuring no recurrence of the issue," Microsoft said, adding it would publish a more detailed review in the next two weeks.
Time to diversify?
With two major outages at US hyperscalers impacting organizations around the world, it's no surprise that questions are being raised about concentrating too much in the hands of a few American tech providers.
AWS holds about 30% of the market, for example, while Azure commands a 20% share, underlining the tight grip both these companies have on the broader cloud computing landscape.
Mark Boost, CEO of Civo, said this latest incident should serve as a warning about diversification within the industry.
"Two of the world’s biggest cloud providers have suffered major outages in the space of a week," he said.
"It’s a wake-up call for governments and enterprises alike, why are so many critical UK institutions, from HMRC to major banks and airports, reliant on infrastructure hosted thousands of miles away?"
Nicky Stewart, Senior Advisor to the Open Cloud Coalition, echoed Boost’s comments on diversification, adding that interoperability and resilience are two key issues that need to be addressed.
"Successive outages on this scale show how a single technical fault can ripple through essential services, public infrastructure, and the wider economy," said Stewart.
"The pattern of repeated disruption underlines the urgent need for diversification, and the CMA must move at pace to implement remedies that foster a more open, competitive, and interoperable cloud market, one where resilience comes from choice, not dependence on the two dominant providers.”
Make sure to follow ITPro on Google News to keep tabs on all our latest news, analysis, and reviews.
MORE FROM ITPRO
- Downtime costs enterprises $400 billion per year
- Manufacturers report millions in losses as downtime wreaks havoc on operations
- Website downtime costs businesses thousands a month
-
Why is big tech rolling back DEI initiatives?In-depth Diverse hiring targets and references to DEI in public documents have fallen markedly over the course of 2025
-
ePrivate 5G and partner ecosystems: The blueprint for intelligent infrastructureIndustry Insights Private 5G networks and collaboration between technology partners underpin the next wave of innovation and enterprise digital transformation
-
Google just confirmed the location of its first small modular reactorNews Developed by Kairos, Google's first small modular reactor will be located in Tennessee, with operations beginning in 2030.
-
Microsoft's plan to use human waste to offset AI emissions stinks of greenwashingOpinion Hyperscalers are getting increasingly gimmicky when it comes to sustainability
-
Microsoft invests $400 million to expand Swiss data centersNews Growing AI and cloud demand, plus data sovereignty requirements, are fueling European data center investment
-
Microsoft says this data center cooling technique can cut emissions by one-fifth – but switching to renewables will prove far more impactfulNews In a cradle-to-grave environmental assessment, Microsoft has examined manufacturing, transportation, deployment, and end-of-life disposal
-
First Microsoft, now AWS: Why tech giants are hitting the brakes on costly data center plansNews Amazon Web Services (AWS) has paused plans for some data center leases, according to analysts, sparking further concerns about the cost of AI infrastructure spending plans.
-
OpenAI inks $12bn CoreWeave deal in latest move away from MicrosoftNews Cloud infrastructure company CoreWeave will supply OpenAI with infrastructure to run the firm's latest models in a deal worth nearly $12 billion.
-
Analysts think Microsoft's data center rollback is bad news for the AI boom – but the company says not to worryNews Microsoft has reportedly ended leases for a significant amount of data center capacity, sparking debate over whether the AI boom is starting to falter.
-
Microsoft invests $700 million to bolster cybersecurity and infrastructure in PolandNews Microsoft has announced plans to invest more than $700 million to support AI and cloud infrastructure expansion in Poland alongside cybersecurity support.
