A power failure in one London zone took Google Cloud customers offline for eight hours
A power event in Google Cloud's europe-west2-b zone on 20 October took services at Revolut, HSBC, Lloyds and Royal Mail offline for most of a working day. Google Compute Engine is now listed here with four European peers.
Google Cloud lost one of its three London availability zones on Tuesday, 20 October 2026. Its status page dates the start to 06:08 UTC, when zone europe-west2-b began failing, and Google later named the cause as “a power event”. Service was restored for most customers by 14:03 UTC, about eight hours later. Google says it will publish an analysis once its internal investigation is complete.
The incident listed 38 affected Google Cloud products, among them Compute Engine, Kubernetes Engine, Cloud SQL, Cloud Storage and BigQuery. The Register and heise reported outages at customers including Revolut, Lloyds Bank, HSBC, Royal Mail, Rightmove, DPD, BT and Screwfix. A same-day outage at ING was unrelated, the bank told heise.
A zone is a separate site with its own power, and a region’s three zones exist so that one can fail without taking the others with it. Google’s advice during the incident was to move affected workloads to another zone or region. Services that ran in europe-west2-b alone had nowhere to go.
This site now lists Google Compute Engine as an incumbent, with four European alternatives that answer the zone question differently. Scaleway Compute Instances runs four regions — Paris, Amsterdam, Warsaw and Milan — each with three zones. OVHcloud Public Cloud Instances has three-zone regions in Paris and Milan; its London region, like most of its others, is a single zone. STACKIT Compute Engine offers a Metro zone that restarts a VM elsewhere in the region if its own zone fails, though its service certificate notes that several zones can share one building. Hetzner Cloud live-migrates servers during maintenance but has no zones inside a location, so redundancy means spreading across its German and Finnish sites yourself.
European clouds fail too. OVHcloud’s SBG2 datacentre in Strasbourg was destroyed by fire in March 2021, and in 2023 a French court ordered OVHcloud to pay damages to two customers whose backups were lost with it. Changing provider changes the jurisdiction a workload runs under; designing across zones is still the customer’s job.