Reality Debunks Tech Myths On Its Own: A Neutral View At A Cloud Outage

October 2025

Earlier today, when I read the news about the AWS outage that affected plenty of companies including several UK banks, I wondered whether I should create a post about this. There are plenty of articles and blog posts already – some are insightful, some are not; some are well informed, others are not. Let me try and add my few cents nevertheless.

To start with: I don’t believe this is an issue that AWS faces alone. We need to keep in mind that many major providers had similar issues in the recent past, but I did not perceive them as publicly visible the AWS outage. Microsoft Azure faced a global outage a bit more than a week ago; and Google Coud had an hourslong outage in June. These few examples indicate that larger outages are not only a AWS issue. From here, let’s step a few steps back and change course slightly.  

In the past, banks and other financial services firms used to have comprehensive disaster recovery strategies with backup data center(s), typically located a significant distance away from the primary data center. There were backup storage and physically separated network cables. Training exercises showed how rapidly and how seamlessly an institution was able to switch to a secondary data center and move back to the primary one.

Availability goals were 99.95% and higher; and objectives were even more demanding in areas like communication networks. Most institutions lived with the fact that high availability and resilience did come with a price tag. Today, some executives still prioritize investing in high availability and resilience, while others don’t.

In the past, there were outages nevertheless (don't tempt me to talk about the almost ancient three-banks disaster). As a result, some banks had to ask their customers to be patient during certain outages.  Today, a UK bank asked its customer to “bear with us”. The difference? In the past bank branches were plenty: for example, when the data center of a European bank went down for a several days many years ago, its customers could still withdraw cash from nearby branches if necessary.

Doesn't it sound as if resilience and disaster recovery strategies should be at least as important today as they were in the past? When debunking the top five myths about banking and insurance tech in last week’s webinar, the Angry Rabbit Group also discussed resilience and high availability in the cloud. The webinar mentioned an positive example: A bank COO in the UK explaining his current preference for a dual-cloud-provider approach. Interestingly, his bank wasn’t one of those mentioned as a victim of the AWS outage today.

What are the lessons learned? Don’t worry: I won’t suggest to increase the number of bank branches. However, here are a few of the lessons that can be learned (they are definitely more): Financial services institutions should                                          

• Build and select application software that can work with at least two different cloud provides (not necessarily hyperscalers).
• Be more demanding when it comes to inhouse application architectures and off-the-shelf application providers they work with.
• Work with providers focusing on different geographies, e.g. a European and a US provider to reduce geo risk, for example.
• Remember good and best practices from the past when their businesses depended on their own data centers only.