On Monday 20 October, shortly after 8 a.m. Central European time, people trying to pay for their morning coffee with a Revolut card saw the payment declined. It wasn’t the card, and it wasn’t the café’s terminal. Within an hour Downdetector had more than 1,500 reports and the app was showing generic errors: sometimes you could log in, but no operation would go through.
The cause, as became clear during the morning, wasn’t inside Revolut at all. It was in a building in London.
Timeline (CEST, UTC+2)
- around 08:00: first reports of declined payments and an unusable app.
- 08:08: the official start of the incident according to Google Cloud’s later account (06:08 UTC).
- around 09:00: Downdetector passes 1,500 reports.
- 09:47: Revolut’s status page lists the incident across many features.
- 14:44: Google says most services are starting to recover.
- 16:03: Google declares the problem resolved “for the majority of affected customers”. Recovery for Revolut users was gradual.
What didn’t work
More or less everything you do with an account: card payments, physical and via Google Pay and Apple Pay, transfers, ATM withdrawals, top-ups, currency exchange, crypto and some investment features. Business customers also lost the web interface, outgoing transfers and the APIs, the automatic connections to accounting tools and online shops. In Italy, Corriere della Sera counted 31 services partially unavailable.
The bank acknowledged the problem and said customers’ money and data were safe. That was true: it was an availability problem, not a security one.
The cause: a Google Cloud “zone” without power
Google reported a power problem in the europe-west2-b zone, one of the three zones of its London cloud region. To see why that matters, a bit of context.
Large cloud providers split each region into several zones: groups of data centres with independent power, cooling and networking. The idea is that if one zone goes down, well-designed services keep running in the others. In theory, a failure in one zone shouldn’t stop anyone.
In practice it often does, for very concrete reasons. Some components, typically primary databases, run in a single zone with a standby copy in another that has to be activated. If failover isn’t automatic, or if the other zones are swamped by every customer moving at once, you get what happened on 20 October. And Revolut wasn’t alone: in the UK, Lloyds, Halifax, HSBC, Royal Mail, DPD, the property portal Rightmove and some UK justice services also had problems.
How to be ready next time
- Keep a second card from another bank, ideally on a different network (a Visa and a Mastercard, for example). If Revolut is your main account, this matters twice as much.
- Keep some cash at home and in your wallet. It sounds old-fashioned, but on 20 October it saved plenty of breakfasts.
- Don’t retry a payment ten times. During outages some transactions show as declined but leave an amount “pending” on the account. Retrying can cause several temporary holds that are reversed only later.
- Check official sources before uninstalling the app or resetting your phone: the service’s status page and its official social accounts.
- Be wary of “refund” or “verify your account” messages in the hours that follow. Scammers always exploit well-known outages with texts and emails that look like they come from the bank. No bank asks for passwords or codes to “reactivate” an account after an outage.
For businesses that take payments through a single provider’s API, the lesson is even clearer: it’s worth having a second payment method configured and ready, even if you use it twice a year.
Source: PagamentiDigitali.it.