Google Cloud's London outage: what it means for your business

On Tuesday 20 October, a power failure in part of Google's London cloud region knocked banking apps, Royal Mail bookings and the courts' video hearings offline for most of the working day.

Security newsPublished By the Expect Best security team4 min read

You may never have signed up for Google Cloud. Your bank, your payment provider and your web host might have.

What happened

At 06:08 GMT (7:08am UK time) on Tuesday 20 October, Google reported a “power event” in its London cloud region. Services running in one zone, europe-west2-b, started returning errors and dropping connections (Euronews).

Google said it had the problem under control just after midday. Services began recovering by 1pm, and by 3pm Google said a very small number of customers still had problems (Silicon UK).

Google has promised a full analysis once its internal investigation finishes. Some posts online claim the backup generators failed. Google has not confirmed that, so treat it as unverified until the official report lands.

Who was affected

The outage hit services that millions of people in the UK use every day:

  • Revolut customers could not pay by card, withdraw cash or make transfers. By 5pm, six of its 27 services were back to normal.
  • Customers of Lloyds, Halifax and Bank of Scotland reported problems with apps and online banking. HSBC said customers had trouble accessing its services.
  • The video system for remote court hearings in England and Wales failed, and reports say several hearings ran late.
  • Customers reported problems with online booking for Royal Mail and DPD, and Rightmove appeared among hundreds of problem reports on Downdetector.

Revolut told customers their money and data stayed safe (Euronews). No report so far points to an attack or a data breach. Customers lost access to services, and their data stayed where it was.

How one zone took down household names

Google splits its London region into three zones, built so that a failure in one leaves the other two running. Companies that spread their systems across all three can lose a zone and keep running on the other two.

That design protects you only if you use it. Any service that ran in a single zone, or kept one essential piece there, such as a database, went down with it. Building across zones costs more and takes more engineering, so plenty of systems skip it.

Your web host makes the decisions about zones. Your part as a business owner is knowing which outside services you rely on, and what you will do when one of them stops.

Was your website affected?

Several UK hosting companies run their servers on large cloud platforms, Google Cloud included. If your website went down on Tuesday morning, ask your host where your site runs and what happens when that data centre fails.

Your site may have stayed up while the services around it failed. Look at the parts of your business that run through someone else’s systems: card payments, online booking, email delivery and your business bank account.

Five things to do this week

  1. List what your business depends on online. Your web host, domain registrar, email provider, payment processor, booking system and bank. Note the support phone number for each one.
  2. Ask your host two questions. Which data centre does my site run in? What happens to my site if that data centre loses power?
  3. Set up uptime monitoring. A free monitor checks your site every few minutes and emails you when it stops responding, so you hear about an outage before your customers tell you.
  4. Plan a fallback for taking money and bookings. Keep a phone number on your site, a card reader that can work offline, and a short social media post ready to publish.
  5. Keep a recent backup away from your host. If your host has a bad day, a copy on separate storage gets you back online faster. Our small business security checklist covers backups in more detail.

Security covers availability as well as hackers. An afternoon offline costs you sales and trust whether a burglar or a power cut caused it.

Common questions

Was the Google Cloud outage a cyber attack?
Google blamed a power event in its London region. No report so far suggests an attack or a data breach.
How long did the outage last?
Google reported the problem at 06:08 GMT on 20 October and said it had it under control just after midday. Most customers had service back by 3pm, with a small number affected for longer.
Was my data at risk?
The reported effects all involved access: people could not reach services. Revolut said its customers' money and data stayed safe. If one of your providers tells you otherwise, follow their advice.
How do I find out if my website runs on Google Cloud?
Ask your web host. A good host will tell you which cloud provider and data centre your site runs in, and how it handles a failure there.

Sources

How safe is your website?

A free A to E grade in about 10 seconds. No account needed.

Check my site