Massive Google outage crashes banking apps: the cash and numbers to keep offline at home

What went down, and when

Google's own incident page is the record. At 07:08 UK time on Tuesday 20 October, Google Cloud began reporting that multiple services were experiencing issues in its europe-west2 region. Its regions documentation gives europe-west2 one location, London, England, with three zones, a, b and c; it is Google Cloud's only region in the United Kingdom. The page's times are shown in US Pacific time, eight hours behind the UK today.

Google's first post, at 09:35, offered no workaround; from 10:24 it advised customers, if feasible, to move affected workloads to another zone or region. At 12:12 it named the cause, a power event in the europe-west2-b zone, and said in the same update that the power event was mitigated, with partial recovery of infrastructure expected within 30 minutes. At 12:58 it said services were recovering, with existing workloads prioritised and new ones possibly still blocked. By 13:10 it was seeing recovery for most workloads, and at 15:03 it posted its last update: the issue was mitigated for most customers, a very small number might still be affected, and an analysis would be published once its internal investigation was complete. The incident is shown as closed at 15:00, seven hours and 52 minutes after it began.

The page lists 38 affected products, among them Compute Engine, Kubernetes Engine, Cloud SQL, Cloud Storage, Cloud Run, Cloud Load Balancing, Pub/Sub and BigQuery: the building blocks that other companies' apps and websites run on. Every update carried the same line: "We apologize to all who are affected by the disruption."

Who said they were hit, and who was only reported

Two companies put their own names to it. Revolut said on X that it was "experiencing a service disruption affecting multiple services", that "all our customers' money and data are safe", and that it had been working on the fault since it was identified in the morning; it listed card payments, transfers, cash withdrawals and sign-in as down or partly unavailable. The outage tracker Downdetector logged more than 1,500 Revolut reports, 1,458 of them by 08:02. The budgeting app Emma said it was "down due to a Google Cloud outage affecting services in London since 7am", that it was a technical outage, not a security issue, and by about 2pm that it was up and running again with data and money safe throughout.

Halifax told a customer on X it was "aware some customers are having issues with our app and online", added that "you can still call us", and did not say whether Google was the cause. Reports of problems at Lloyds and Bank of Scotland, two other brands in the same banking group, and at HSBC, Royal Mail, DPD and Rightmove came from Downdetector users, not from the companies: by 09:00 Downdetector held 480 reports about the Lloyds app and website, the largest share, 36%, about logging in. None of those companies had published a statement that we could find, and it is not known whether all of them use Google Cloud.

In the courts, the Criminal Bar Association said on X that the Cloud Video Platform was "down across large parts of England and Wales since 10am" and that prosecution and defence counsel "can not proceed with trials and other hearings". HM Courts and Tribunals Service confirmed it was "aware of an issue affecting Cloud Video Platform, used for some remote hearings". Journalists using the platform told the Press Association it resumed shortly after Google's noon update.

The Daily Mail's headline also named the NHS. We could find no statement from NHS England by Tuesday evening, so we do not count it. Nobody has published how many people were affected in total.

Why one power fault reaches your bank

Google Cloud rents out computing and storage from large data centres. A region is a geographic area with several zones; Google describes a zone as a deployment area within a region, designed so that a failure in one is less likely to spread to the others. Its advice from 10:24, to move affected workloads to another zone or region, is that design in use: a company running across more than one zone is less exposed, a company with everything in europe-west2-b is not.

Which of Tuesday's affected services sat where is not public, and we make no guess. What the day shows is simpler: several services that people treat as separate, a bank, a budgeting app, the courts' video link, can share one supplier, one city and, on a bad morning, one power fault.

What to keep offline at home

Our household kit list has carried the same line since the site began: some cash, because card machines and ATMs fail in a power cut. Tuesday was not a power cut at home, but the effect for a Revolut customer at a till was the same, and the fix is the same. A modest amount of small notes covers a day's essentials. When the card machine's down sets out how much and where to keep it.

A second card from a different banking group matters more than a second card from the same one. Halifax, Lloyds and Bank of Scotland are brands of one group; two cards from two of them share more than two cards from unrelated banks. Which? Money's Sam Richardson, quoted by the Press Association on Tuesday, recommended the same two things: a card with a different bank and a little emergency cash. The same goes for apps: if your budgeting app, your bank and your backup bank all run on the same cloud region, they fail together.

Write down the phone numbers you would need without the app: your bank's lost-and-stolen and customer lines (Halifax's own advice on Tuesday was that customers could still call), and anything else you only ever reach through a login. Keep a paper copy of the day's essentials in the same place as the cash. Our emergency kit list and the kit builder put these next to the torch and the radio, which is where they belong.

If a failed payment on Tuesday cost you money, a late fee, a parking charge, a missed transfer, tell your bank first and keep the times. If the bank does not put it right, the Financial Ombudsman Service handles complaints about banks and is free to use. That is information and the official route, not advice on your case.

What happens next

Google has said it will publish an analysis of the incident once its internal investigation is complete. We will add what it says about the power event and the recovery to this page when it appears, and the time at the top will change. Our guide to a national power outage covers the wider case, where the fault is at the grid rather than in one data hall, and the landline switch-off explains why a digital phone line needs power too.