Have Cloud Outages Become Too Common to be Newsworthy?
On Tuesday, March 18 Microsoft Azure had an hours-long significant and widespread service interruption. Large government agencies and companies were out of commission for various lengths of time. Public cloud outages have become so common that in the 24 hours that followed it didn’t even make the news.
Now this is not an anti-cloud post. Nearly all organizations utilize host applications in the most appropriate environment for the specific requirements of each. A hybrid blend of some in public cloud and some on data center infrastructure in their own data center or within their footprint in a colocation data center has become the most common choice.
While most company’s bread and butter IT works best in traditional data center infrastructure there are other applications, like global e-commerce or video delivery to a wide geographic footprint, that can thrive in AWS, Microsoft, Azure, Google Cloud, and other large hyperscale providers that can be worth the increased cost for those users. Some internal businesses processes, like email, accounting, some HR, etc. also make sense outsourcing to the cloud.
But the complexity of these public cloud networks, as impressive as they are, also leads to service interruptions. The March 18 Microsoft Azure outage was not an outlier. There was another Azure outage in early March that hit tens of thousands of users and others so far in 2025 that are detailed here, here, and here. Service interruptions within AWS and other cloud providers also occur regularly, industry-wide averaging 77 per week in the first two months of 2025, as we showed in a recent blog post.
Despite cloud providers’ best efforts to avoid outages, experience shows that they are inevitable. Executing on a sensible disaster recovery strategy when the occur should be a priority for you if your organization is in the public cloud.
Preparing for Cloud Outages: Are you prepared for when your applications in the public cloud go down due to service provider outages? That’s an issue addressed in the Direct LTx executive report Maintaining High Availability: 9 Critical Steps to Take for Disaster Recovery Success, which you can read here.