Source: ABC News

A “power event” at a Google Cloud data centre in London disrupted cloud services across Europe for hours on 20 October 2026. Google attributed the failure to the europe-west2-b availability zone inside its London region, and the effects rippled far beyond one building. Major UK banks including Revolut reported service disruptions, customers struggled to reach mobile banking, court live streams across Britain failed, and websites across the continent showed elevated errors and intermittent network issues. Google said the incident began shortly after 6am UK time and was largely resolved through the afternoon, followed by an internal investigation.

One zone, many broken services

The uncomfortable lesson from this event is how much of UK finance and public services now depends on a single cloud region, and in some cases on a single zone within it. An availability zone is supposed to be an isolated failure domain, but as the incident shows, a power event in one zone can cascade into service errors across the wider region when applications, databases and identity services were all placed there without a fallback.

Resilience is a design decision, not a subscription feature. Every serious cloud architecture should spread critical workloads across multiple availability zones as a baseline, and consider a second region or a second provider for anything that genuinely cannot tolerate hours of downtime. That includes asking hard questions about databases, queues, DNS and CDN failover, because an application that survives a zone loss only on paper is not resilient in practice.

If your organisation runs critical workloads on Google Cloud, AWS, Azure or any European provider and you are not certain you would survive a regional power event, contact Excello Digital at https://excello.digital/contact/. We design and review cloud architectures for high availability, run failure-mode workshops and build automated failover that is actually tested before you need it.