Skip to main content

Kaer NewsSoftware

Google Cloud power event knocks out London zone and UK bank services

A power event in Google Cloud's europe-west2-b zone began at 0608 UTC on 20 October 2026, disrupting Lloyds, HSBC, Royal Mail and Revolut, with full recovery tentatively due by 1300 UTC.

4 min read
Google Cloud power event knocks out London zone and UK bank servicesPhoto: The Register

Key points

A power event in Google Cloud's europe-west2-b London zone starting at 0608 UTC on 20 October 2026 disrupted UK banks, Royal Mail, courts and the NHS app.

Google Cloud's europe-west2-b zone in London failed after what the company called a power event, disrupting services for customers across the United Kingdom from 0608 UTC on 20 October 2026. Banks including Lloyds and HSBC, parcel firm Royal Mail, property site Rightmove and fintech Revolut reported problems. Google's status page said at 1158 UTC that full recovery of workloads was tentatively due by 1300 UTC.

The outage matters because it shows how a single cloud availability zone failure cascades through dependent services across an entire digital economy. Europe-west2-b is one of three availability zones in Google's London region, designed to protect against hardware or regional failures. Dr Amro Al-Said Ahmad, a lecturer in computer science at Keele University, estimated 38 services were affected, showing how failures propagate through applications, APIs, databases, payment platforms and third-party software products.

How one zone failed

Google Cloud attributed the outage to a power event in the europe-west2-b zone, stating on its status page: "A power event caused an outage within europe-west2-b zone. The power event is mitigated." The company said it was seeing recovery of services and prioritising existing workloads, though customers might not yet be able to create new workloads. Google had not offered further explanation beyond its status page when asked.

The chain of failure runs from the power event through the zone's infrastructure to the services running on it, and then to the customer applications that depend on those services. Google's status page reported elevated errors and intermittent network failures across the europe-west2 region. Dr Ahmad explained that when a core cloud region fails, applications, APIs, databases, authentication systems, payment platforms and analytics pipelines that depend on it experience cascading failures.

Google Cloud listed the affected services as BigQuery Data TransferService, a tool that moves data into Google's analytics warehouse; Cloud Filestore, its file storage; Cloud Run and Google App Engine, its application hosting platforms; Google Cloud NetApp Volumes; Google Kubernetes Engine, its container orchestration service; and Managed Service for Apache Kafka, its data streaming service. Google told customers there were no workarounds and suggested migrating impacted workloads to another available location if feasible.

Which services went down

Downdetector reports identified affected customers including Lloyds Bank, HSBC, Royal Mail, Rightmove, DPD, BT, Screwfix and Revolut. Revolut confirmed a service disruption affecting multiple services and said all customers' money and data were safe. Lloyds said its app was working but acknowledged some people were having issues with certain online services. HM Courts and Tribunal Services said it was aware of an issue affecting its Cloud Video Platform for remote video hearings, and patients reported glitches on the NHS app.

The disruption was less severe than a comparable outage the previous year, when Amazon Web Services went down for about 15 hours, hitting us-east1 in Virginia, the densest concentration of datacentres on Earth. That outage took out Roblox and Signal and caused smart mattresses to heat overnight. The Amazon outage was later attributed to a bug in automation software, not a power failure. Today's problems peaked around 9am and were somewhat resolved by mid-afternoon, per Downdetector reports.

How it compares to AWS

Michał 'rysiek' Woźniak, an internet infrastructure expert, said the outages showed the concentration of vital services on a few key companies he called hyperscalers. He said Google's engineers were doing what they could, but argued it was unreasonable for a handful of gigantic companies to cause public institutions and large parts of the digital economy to stop working. He also noted Google had said 75% of new code was AI-generated and approved by engineers.

The scale of cloud dependence in Britain raises the stakes of such failures. Up to 60% of UK government infrastructure is hosted on cloud services, mostly Amazon, Microsoft and Google, and British police keep vital data on Microsoft's Azure cloud. Dr Ahmad said resilience is not achieved simply by moving to the cloud, and that true resilience requires designing for regional failures through multi-region deployments, workload portability and tested disaster recovery procedures.

Google Cloud's status page update at 1158 UTC on 20 October 2026 said full recovery workloads should be completed tentatively by 1300 UTC, with customers to expect another update at that point. The company had earlier planned an update around 0930 UTC and told customers there were no workarounds for the time being. Google's engineers continued to investigate the issue, and the company apologised to those affected by the disruption.

Frequently asked questions

What caused the Google Cloud London outage?

Google Cloud said a power event caused an outage in its europe-west2-b zone in London. The company stated the power event was mitigated and services were recovering, though it offered no further explanation beyond its status page.

Which UK services were affected by the outage?

Downdetector reports identified Lloyds Bank, HSBC, Royal Mail, Rightmove, DPD, BT, Screwfix and Revolut as affected. HM Courts and Tribunal Services reported issues with its Cloud Video Platform, and patients reported glitches on the NHS app.

When did the outage start and when was recovery expected?

The downtime began at 0608 UTC on 20 October 2026. Google's status page said at 1158 UTC that full recovery workloads should be completed tentatively by 1300 UTC, with another update expected then.

More from this edition