Hosting
The Google Cloud outage on October 20, 2026 knocked banking apps, parcel tracking and court video links offline across the UK for most of a working morning. Google says “a power event caused an outage within europe-west2-b zone,” one of three zones in its London region, The Register reports. Trouble started at 06:08 UTC and Google called the issue mitigated for most customers at 14:03 UTC. Our view is plain: a famous cloud name on the invoice tells you very little about uptime. What protects a website is how it is built to survive a failure and how fast it can come back.
This post covers what failed, why one building’s power problem spread so far, what it means if your WordPress site runs on a cloud host, and the questions worth asking your provider this week.
What Happened in the Google Cloud Outage
Google’s status page first reported “elevated errors and intermittent network failures in europe-west2 region,” Tom’s Guide reports. Later updates blamed a power event in a single London zone. Google said it was “prioritizing the recovery of existing workloads” and warned that customers “may not be able to yet create new workloads,” per The Register.
Google’s preliminary report gives more detail. According to IncidentHub’s summary of that report, the datacenter behind europe-west2-b lost power, and both the primary and backup controllers that should have started the emergency generators failed to do so. Battery backup then ran down, and servers, network gear and storage in the zone went offline. IncidentHub says traffic that failed over to the other two London zones then overloaded parts of them, which is why customers outside the failed zone also saw errors. Google has said the report is subject to change and that a final version with preventive actions will follow.
Key facts at a glance
- Date: Monday, October 20, 2026.
- Region: europe-west2 (London), with the failure starting in zone europe-west2-b.
- Cause: a power event, per Google’s status page as quoted by The Register and ABC News.
- Length: about six hours of core impact in Google’s preliminary timeline, with the incident closed at 14:03 UTC.
- Google products hit: The Register lists Cloud Run, App Engine, Google Kubernetes Engine, Cloud Filestore and others. IncidentHub adds Compute Engine, Cloud SQL, Cloud Storage and BigQuery.
Who Went Down and Why It Spread
The list of affected brands reads like a tour of daily life in Britain. Revolut, Lloyds Bank and Halifax customers reported app and online banking problems, according to TechRadar and ABC News. DPD, Royal Mail tracking and Rightmove also had trouble, per The Register and Tom’s Guide. ABC News and Tom’s Guide both report that HM Courts and Tribunals Service had problems with the video platform used for some remote hearings.
Not every affected company ran its own systems in the failed zone. IncidentHub notes that multi-zone and regional services also saw higher latency and errors. Modern apps also lean on databases, payment providers, login services and APIs, and when one of those pieces sits in a failed zone, everything built on it can stall. IncidentHub tracked 12 downstream outages at other providers, including Stripe, Snowflake, MongoDB and the WordPress host Kinsta. Each of those has its own customers, so a single power failure became a chain of failures.
Dr Amro Al-Said Ahmad of Keele University put it well in comments to The Register: “resilience is not achieved by simply moving to the cloud.” He said “true resilience requires designing for regional failures” through multi-region deployments, workload portability and tested disaster recovery.
The cloud behaved the way its own documentation says it can, with a single zone failing. The sites that stayed up were the ones built to expect that.
The Debate: Does Big Cloud Mean Better Uptime?
Many hosts advertise that they run on Google Cloud, AWS or Azure, and the pitch implies that a giant provider means your site will not go down. There is a real case for that view. The big clouds have strong security teams, huge networks and more redundancy in each building than a small company could afford. For most hours of most days, they are very reliable.
The other side of the argument is that the cloud only gives you the parts. Google itself says that “putting resources in different zones in a region reduces the risk of an infrastructure outage,” and that “putting resources in different regions provides an even higher degree of failure independence,” in its regions and zones documentation. Google also recommends deploying applications across multiple zones and regions to guard against failures “up to and including a single zone or region.” During the incident, Google suggested affected customers consider moving workloads to another zone or region where feasible, TechRadar reports.
In other words, Google’s own guidance assumes a zone can fail. The cloud gives a host the option to spread sites across zones and regions. Whether the host actually does that, and pays for it, is a separate decision. A WordPress site sitting on one virtual machine and one database in one zone is just as exposed on a premium cloud as it would be in a single small data center.
Why most WordPress sites live in one place
Running a typical WordPress site across several zones is harder than it sounds. WordPress keeps its content in a MySQL or MariaDB database and its media in the uploads folder. Keeping both copies in sync across buildings takes database replication, shared or replicated file storage, and a plan for which copy wins when the network splits. That adds cost and complexity. For many small business sites, a single well-run server with strong backups is a sensible trade. The problem starts when the marketing suggests more protection than the setup really provides.
What a Six Hour Outage Costs
Uptime promises are easy to read and easy to misread. Here is the plain math for a 30 day month, which has 43,200 minutes:
- 99.9% uptime allows about 43 minutes of downtime a month.
- 99.95% uptime allows about 22 minutes a month.
- 99.99% uptime allows about 4 minutes a month.
Six hours is 360 minutes. A site that was down for the full window on October 20 would land near 99.2% for the month, well short of any of those targets. A service credit under an uptime guarantee is usually a slice of the monthly fee, which rarely matches the orders, leads or bookings lost in a morning.
For a store, a booking site or a lead form, the real cost is the business that does not happen while the page will not load. You can estimate your own number by taking your average revenue per hour during business hours and multiplying it by six. Then add the time your team spends answering confused customers. Our post on managed WordPress hosting infrastructure walks through the building blocks that keep that number small.
What to Ask Your Host After the Google Cloud Outage
You do not need to read an architecture diagram to judge your risk. Ask your host these questions and listen for direct answers.
- Where does my site physically run? Ask for the provider, the region and whether your site, database and files sit in one zone or several.
- What happens if that zone or data center goes dark? A good answer names a real plan, such as failover to another zone, restore to another location, or cached pages served from the edge. “That never happens” is a bad answer, as October 20 showed.
- Where are my backups stored? Backups kept on the same server or in the same zone can disappear in the same event. Off-site copies in a different location are the ones you can restore from. Our guide to WordPress backup solutions explains retention and storage choices.
- How long would a full restore take? Ask for a time in hours and ask when they last tested it.
- How will you tell me what is going on? A public status page, email alerts and a human on the support line matter most in the first hour.
Use your CDN as a safety net
A content delivery network can keep part of your site visible during an origin outage. Cloudflare’s Always Online feature serves cached or archived copies of static pages when Cloudflare cannot connect to your server at all. It has limits. Cloudflare says it “cannot serve dynamic content while your origin is offline,” so carts, checkouts, logins and forms still fail. It also does not save every page, and it does not kick in when your server is reachable but returning errors. Even so, a visitor who sees your homepage and contact details is better off than one who sees a blank error screen.
Load balancing across servers is the next step up for sites that cannot afford downtime. Our post on WordPress hosting load balancing covers how traffic can shift between servers when one fails.
How WP Farm Thinks About Outages
No host can promise that a data center will never lose power. We plan for the day something breaks. Every WP Farm plan includes daily off-site backups, so a site can be restored somewhere other than the server it normally lives on. The sites we manage run behind Cloudflare, which gives us edge caching and filtering in front of our servers. When something does go wrong, a real person on our team picks up the ticket and tells you what we see and what we are doing about it.
We would rather tell customers exactly how their site is protected than lean on a famous cloud name. If you want to know where your site runs and how it would come back after a failure like this one, ask us and we will show you.
Google Cloud Outage FAQ
What caused the Google Cloud outage on October 20, 2026?
Google says a power event caused an outage in the europe-west2-b zone in London. Its preliminary report, as summarized by IncidentHub, says the emergency generators failed to start and battery backup ran out. Google plans to publish a final report.
Was my WordPress site affected?
Only if your site, or a service it depends on, ran in Google Cloud’s London region. IncidentHub lists the WordPress host Kinsta among the downstream outages. If your site was slow or down that morning, ask your host where it runs.
Does multi-zone hosting stop every outage?
No. It lowers the risk from a single building failing, and Google recommends it for that reason. A whole region can still have problems, and failover traffic can overload the zones that remain, as IncidentHub described. Good backups and a tested restore plan are still needed.
Is an uptime guarantee worth anything?
It tells you what a host is willing to commit to. The payout is usually a credit on your bill, which is small next to a morning of lost sales. The better test is how the host explains its recovery plan.
Talk to Us
If the Google Cloud outage made you wonder how your own site would hold up, contact our support team. We will tell you plainly where your site runs today, how it is backed up and what a recovery would look like.
Hosting
Hosting
Hosting