A regional cloud failure became a banking outage
Google Cloud reported elevated errors and intermittent network failures across its London region, europe-west2. The official incident record listed failures across a broad infrastructure set, including Compute Engine, Cloud Storage, Kubernetes Engine, Cloud SQL, BigQuery, Cloud Run, Firestore, Pub/Sub and Virtual Private Cloud. This was not one broken application. It was a regional failure affecting compute, storage, databases, networking and managed services at the same time.
The customer-facing consequence was clearest at Revolut. Its official status page marked 30 functions as partially unavailable, including sign-in, card payments, transfers, ATM withdrawals, currency exchange, stock trading, cryptocurrency, business applications and merchant payments. Reuters reported that a person familiar with the matter linked Revolut's disruption to Google Cloud. Revolut said customers' money and data remained safe.
Timeline: trouble overnight reached the checkout
Google's status record says the incident began on October 19 at 11:08 p.m. Pacific time, equivalent to 7:08 a.m. in London on October 20. At 1:35 a.m. Pacific, Google said engineers were investigating and that no workaround was available. By 2:24 a.m., it suggested that affected customers could try moving workloads to another zone or region if feasible.
By 9:11 a.m. UTC, Reuters reported that Revolut's card payments and ATM withdrawals were among the services unavailable or partially unavailable. The Guardian separately reported failures affecting account access, card transactions, cash withdrawals and transfers, with customers encountering errors when opening the app.
That sequence matters. A cloud provider's regional incident moved rapidly from infrastructure telemetry to ordinary people being unable to complete basic financial actions. The technical boundary was a data-center region; the practical boundary extended to shops, cash machines and customer accounts.
What is known, and what is not
The evidence establishes a Google Cloud regional disruption, extensive Revolut service degradation, and a reported link between the two. It does not yet establish the underlying technical cause, the number of failed transactions, the number of customers affected, or whether every listed Revolut function failed for every user.
ING Germany also reported login trouble during the same period, but Reuters said it was unclear whether that problem was related. It should not be counted as part of the Google Cloud incident without further evidence. Google had not published a root-cause analysis or final restoration time when this draft was prepared.
Resilience has to follow the customer journey
The immediate lesson is that choosing a cloud region is a deployment decision, not a continuity plan. A service is resilient only when critical workloads can move, their data remains consistent, dependencies are available elsewhere, and operators have rehearsed the change under realistic conditions. A suggestion to migrate during an active incident is not equivalent to a tested failover.
Financial-service testing should follow complete user journeys. Separate green checks for an app, a card processor and a transfer service can miss their shared dependency. A useful exercise starts with outcomes: can a customer sign in, see an accurate balance, pay at a shop, transfer money and withdraw cash when the primary region is unavailable?
The Revolut status page also shows the danger of correlated failure. Access to the app, cards, transfers and cash withdrawals degraded together. Customers may reasonably treat those as alternate ways to reach their money, but shared infrastructure can make the alternatives fail at once. Continuity design should preserve at least one independently operated route for essential transactions, and communications should tell customers which actions actually work.
What remains uncertain
The incident was ongoing when researched. Google had not published a root cause, full impact metrics or a final recovery time. Revolut described services as partially unavailable, so the draft does not assume universal failure. Reuters linked Revolut to the Google Cloud incident through a knowledgeable source; Revolut's public status page did not name Google. ING Germany's simultaneous problem is excluded from the incident scope because a connection was unconfirmed.
Sources & method
AI-assisted research and drafting, reviewed by the editor before publication. Reporting is attributed to the sources below; this is not an independent forensic investigation.
- Multiple services are experiencing issues in the europe-west2 region · Original source · 2026-10-20
- Revolut system status · Original source · 2026-10-20
- Google Cloud services face outage in London, Revolut services down too · Reporting · 2026-10-20
- Millions unable to use Revolut app or cards as company suffers service disruption · Reporting · 2026-10-20