Dispatch, 2026-11-12, SHC Team, Cloud dependence
A power fault took down a Google Cloud zone in London
Filed 2026-11-12 · SHC Team
Early on Oct. 20, London time, a power fault cut electricity to one zone of Google Cloud's London region. Google's own incident report puts the disruption across the region at 6 hours and 8 minutes. The banking app Revolut told customers that "a recent Google Cloud outage" had disrupted its app that day.
What happened
A cloud region is split into zones, separate data center sites, and Google says it designs regional services to survive losing one. The site hosting zone europe-west2-b was down to one of its two power feeds. The other had been switched off for repairs after a generator test on Oct. 14 turned up problems. When an underground cable fault outside the building cut the remaining feed, the backup generators did not start. The control system read live voltage on the feed under repair and sent no start signal. Batteries that last about five minutes ran out, and the data halls lost power.
Power returned that morning, but storage that had shut down suddenly needed a long recovery before services restarted. Some customers in London's other two zones were hit too. Failover, the switch to standby systems, swamped spare capacity, and some failover attempts failed. One Google service, Apigee X, is built to survive a zone loss, yet two of its three database copies sat in the zone that lost power.
A different fault hit Google Cloud's Iowa region, us-central1, on Sept. 1. During routine router maintenance, a technician was handed every router's changes at once instead of one router at a time, and unplugged all fiber paths within 13 minutes. Some customers were cut off or slowed for 4 hours and 11 minutes.
What changes for you
Neither outage was an attack. In both, a fault at one site cut off the customers whose systems ran only there. Google's report is plain about who covers that risk: for servers in a single zone, customers must set up their own failover and recovery. And failover itself can fail when everyone switches at once.
For a small business, law firm, farm or nonprofit, the exposure is usually the apps you rely on: your bank, payroll, bookings, email and files may each sit with one provider in one place. You can't redesign their systems. You can know where your records live, keep a copy that doesn't share the same weak point, and plan for a day without them.
What you can do
- Know where your records live. Ask whoever handles your tech, and each key vendor, which provider and region hold your email, files and records, and whether they can keep running from a second zone or region.
- Keep a second copy. Keep what you can't lose somewhere else as well, with a different provider or on a drive you control.
- Test a restore. Bring a file back from the backup and check that it opens.
- Plan for a day without it. Write down how you would take payments, reach customers and find key records if your main app were down for a day.
- One option: SHC runs backup and recovery for email, shared files and laptops, and restores a file you pick so you can check it opens. No one can promise you'll never lose anything, SHC included.
Sources
- Google Cloud Service Health, "Multiple services were experiencing issues in the europe-west2 region." (Incident Report), 2026-11-04
- Revolut (@Revolut on X), "Due to a recent Google Cloud outage, customers may have experienced temporary service disruption on the Revolut app today.", 2026-10-20
- Google Cloud Service Health, "Multiple products in us-central1-b are experiencing network service degradation." (Incident Report), 2026-09-10