Google Cloud experienced a major service disruption in its London region, europe-west2, on 20 October 2026. Google reported elevated errors and intermittent network failures across multiple services, while Revolut confirmed that several customer-facing features were unavailable or partially unavailable.cloud.google+1
For developers, this incident is a clear reminder that even a highly available cloud platform can fail — and that application architecture needs to assume that failure will happen.
What happened with Google Cloud?
Google Cloud’s official status page reported that multiple services were experiencing issues in the europe-west2 region, its London data-centre region. The incident began at 23:08 PDT on 19 October 2026, equivalent to 07:08 Irish time on 20 October 2026.cloud.google+1
Google described the symptoms as elevated errors and intermittent network failures. Its engineering team said it was investigating, but no root cause, resolution time, or workaround was initially published.reuters+1
The incident affected services including:
- AlloyDB for PostgreSQL
- Apigee
- BigQuery Data Transfer Service
- Google Kubernetes Engine
- Cloud Filestore
- Cloud Run
- App Engine
- Compute Engine
- Cloud Storage
- Pub/Sub and Dataflow
Why was Revolut down?
Revolut reported a service disruption affecting multiple services, including card payments, bank transfers, ATM withdrawals, currency exchange, identity verification, and live support chat. Its status page showed that several of these services were unavailable or partially unavailable.
A source familiar with the matter told Reuters that Revolut’s outage was related to the Google Cloud disruption in London. Revolut also confirmed that customer money and data were safe.
This illustrates an important point: a cloud outage does not need to affect an application’s code directly to take the product offline. If the infrastructure behind authentication, payment APIs, internal messaging, databases, or service-to-service communication becomes unavailable, users experience failed payments, login errors, and app timeouts.
Which services were affected?
Revolut was the most visible consumer-facing casualty, but reports also linked the disruption to problems at other UK services, including banks and delivery platforms. However, the exact impact on each company may differ, so it is best not to claim that every outage had the same technical cause.reuters+1
What is europe-west2?
europe-west2 is Google Cloud’s London region. Cloud regions are geographic locations containing multiple availability zones, which are physically separate data-centre facilities designed to reduce the blast radius of hardware, networking, or power failures.
In theory, spreading workloads across zones protects an application from a single-zone failure. In practice, many products still have dependencies concentrated in one region because of latency, data-residency requirements, cost, managed-service limitations, or architectural decisions made early in a project.
That is why a regional disruption can still affect companies whose primary infrastructure is in London — even if their application is technically distributed.
What developers can learn from the outage
1. Cloud is not automatically highly available
Using Google Cloud, AWS, or Azure does not guarantee uptime. Availability is an architectural outcome, not a feature you get simply by deploying to a managed cloud.
2. Design for graceful degradation
If a payment provider, database, authentication service, or third-party API fails, your app should degrade rather than break completely. Examples include:
- Showing cached balances or product data where appropriate
- Queueing non-critical actions instead of failing immediately
- Displaying a clear status message instead of a generic 500 error
- Separating critical payment flows from non-critical features such as notifications
3. Use retries carefully
Retries can help with transient network failures, but blind retries can amplify an outage. Use exponential backoff, jitter, request timeouts, circuit breakers, and rate limits to avoid overwhelming a struggling dependency.
4. Monitor dependencies, not just your own app
Your monitoring should cover:
- API latency and error rates
- Database connectivity
- Queue depth
- Third-party provider status
- Frontend error rates
- User-facing transaction success rate
A green deployment does not mean much if a downstream cloud dependency is failing.
5. Have a multi-region plan — but understand the trade-offs
Multi-region architecture can improve resilience, but it introduces complexity around data consistency, cost, latency, failover testing, and managed-service support. For many smaller applications, the first priority is not multi-region; it is clear failure modes, backups, monitoring, and a documented incident-response process.
Is Revolut still down?
At the time of writing, Revolut’s status page still listed an active incident affecting services such as card payments, transfers, and ATM withdrawals. Google Cloud also still listed the europe-west2 incident as under investigation, with no public root cause or final resolution time.cloud.google+1
For the latest status, check:
FAQ
Is Google Cloud down in London?
Google Cloud reported issues affecting multiple services in its London region, europe-west2, on 20 October 2026. The company reported elevated errors and intermittent network failures.cloud.google+1
Is Revolut down because of Google Cloud?
Revolut confirmed a service disruption affecting multiple features, and Reuters reported that a source familiar with the matter linked the outage to the Google Cloud disruption. Google had not publicly confirmed the exact technical relationship at the time of writing.
What is the Google Cloud europe-west2 region?
europe-west2 is Google Cloud’s London region. It hosts cloud infrastructure used by companies across the UK and Europe.
Did customer money or data become unsafe?
Revolut said that customer money and data were safe during the disruption. The incident appears to have affected service availability rather than being a confirmed data breach.
What should developers do during a cloud outage?
Check the provider’s status page, avoid aggressive retries, use exponential backoff, communicate clearly with users, monitor error rates, and activate your incident-response process if critical user flows are affected.
