A power event in Google’s europe-west2-b (London) zone triggered a wide Google Cloud outage on October 20, 2026, taking down dozens of services including Cloud SQL, BigQuery, Compute Engine, and Kubernetes for customers across the UK and beyond. StatusGator sent an Early Warning Signal at 07:02 UTC, 1 hour and 35 minutes before Google Cloud publicly acknowledged the incident at 08:37 UTC. The ripple effect was wide: StatusGator tracked 171 other services showing signs of impact during the outage.
Incident at a glance
| Detail | Value |
|---|---|
| Service | Google Cloud |
| Date | October 20, 2026 |
| Status | Resolved |
| Duration | ~3 hours 45 minutes of user reports |
| Root cause | A power event in the europe-west2-b (London) zone (per Google) |
| Scope | europe-west2-b zone, with wider elevated errors across the europe-west2 region |
| Affected services | Dozens, including Compute Engine, Kubernetes Engine, Cloud SQL, BigQuery, Cloud Storage, Cloud Run, Spanner, Pub/Sub, and Load Balancing |
| First user report | 06:22 UTC |
| StatusGator Early Warning Signal | 07:02 UTC |
| Public acknowledgment | 08:37 UTC |
| Recovery | 10:07 UTC (last user reports) |
| Detection lead | 1 hour 35 minutes before public acknowledgment |
| Downstream services affected | 171 tracked by StatusGator |
Was Google Cloud down on October 20, 2026?
Yes, in one zone. Google Cloud’s europe-west2-b (London) zone suffered a multi-service outage on October 20, 2026, with knock-on effects across the wider europe-west2 region. This was not a global Google Cloud failure. It was concentrated in a single zone, which is exactly why the impact clustered so heavily in the United Kingdom. For customers running workloads there, though, a zonal outage is a full outage: databases, queries, and hosted apps all stopped responding.
The outage was caused by a power event in Google’s europe-west2-b zone. In its incident report, Google said “a power event caused an outage within europe-west2-b zone,” which explains why so many unrelated services failed at once. When a single zone loses power, everything running in it goes down together, from databases to compute to storage.
Google’s updates described services as unavailable in the europe-west2-b zone, with some services in the broader europe-west2 region also seeing elevated errors and intermittent network issues. Google reported the power event as mitigated several hours in, then prioritized recovering existing workloads, while noting that customers might not yet be able to create new ones. Google also stated that Looker Studio was not impacted by the outage.
How long was Google Cloud down?
User reports spanned about three hours and 45 minutes, from the first at 06:22 UTC to the last at 10:07 UTC. Google dated the start of impact earlier, to 06:08 UTC (23:08 PDT on October 19), and its recovery ran longer still: Google reported the power event mitigated and workloads recovering through the late morning and into the early afternoon UTC. Google Cloud acknowledged the incident publicly at 08:37 UTC, partway through the user-report window.
Timeline of the October 20 Google Cloud outage (UTC)
| Time (UTC) | Event |
|---|---|
| 06:08 | Google later dates the start of impact to here, a power event in the europe-west2-b zone (23:08 PDT, Oct 19) |
| 06:22 | First outage report reaches StatusGator, 14 minutes after Google’s stated start |
| 06:22 to 07:01 | Reports build, dominated by connectivity failures and unresponsive services from UK users |
| 07:02 | StatusGator sends an Early Warning Signal to subscribers |
| 07:02 to 08:13 | Reports name Cloud SQL, BigQuery, and hosted apps failing in europe-west2 |
| 08:14 | Google acknowledges the incident through its private status API, 23 minutes before its public page |
| 08:37 | Google Cloud publicly acknowledges the incident on its status page |
| 10:07 | Final outage reports arrive at StatusGator as the zone recovers |
| ~11:12 | Google reports the power event mitigated and workload recovery underway |
Note the 08:14 UTC entry. That is when Google’s private status API reflected the incident, 23 minutes before the public status page updated at 08:37 UTC. StatusGator’s private status integrations pull a provider’s internal, account-specific health data directly from APIs like Google Cloud Personalized Service Health, rather than waiting for the public page. That surfaces a provider’s own acknowledgment sooner, though in this incident StatusGator’s crowd-based Early Warning Signal at 07:02 UTC was earlier still.
Who and what was affected?
Because a whole zone lost power, the blast radius was enormous. Google listed dozens of affected services, spanning compute, storage, databases, networking, and analytics. The ones that showed up most in user reports, and that matter most for everyday workloads, included:
- Compute: Compute Engine and Google Kubernetes Engine
- Databases: Cloud SQL, Cloud Spanner, Firestore, Bigtable, and Memorystore
- Analytics: BigQuery, with queries and scheduled queries failing
- Storage and networking: Cloud Storage, Persistent Disk, Cloud Load Balancing, and VPC
- Hosted applications and third-party platforms running on Google Cloud in the region
Reports were overwhelmingly from the United Kingdom, with a smaller cluster in India and scattered reports from Pakistan, Spain, Romania, South Africa, and beyond. A sample of what users reported, with identifying details removed:
- “Europe west 2b region seems to be totally down” (Lagos, Nigeria)
- “GCloud SQL issues in Europe West 2 and Europe North 2 regions.” (Islington, England)
- “I can’t run BigQuery queries. It comes with error that Google wasn’t able to process my request.” (Nottingham, England)
- “Cloud SQL saying 0 out of XTB. Looks like an outage in eu-west 2.” (Derby, England)
- “Our company job management system running on Google Cloud is no longer available.” (St Helens, England)
- “Queries in BigQuery are not working.” (Kensington, England)
StatusGator’s Google Cloud outage map, showing where reports concentrated on October 20, 2026.
The workaround Google suggested
Throughout the incident, Google advised affected customers to migrate impacted workloads to another available zone or region where feasible. For a zonal power failure, that is the practical escape hatch:
- Fail over to another zone within europe-west2, or to a different region entirely.
- Route traffic away from the affected zone using a load balancer if your architecture already spans zones.
- For data services, promote a replica in a healthy zone or region if one is available.
None of this helps if your workload only ever ran in a single zone, which is the real lesson of a zonal outage: redundancy has to exist before the power goes out, not after.
How did StatusGator detect the outage early?
StatusGator sent its Early Warning Signal at 07:02 UTC, 1 hour and 35 minutes before Google Cloud publicly acknowledged the incident at 08:37 UTC. It works by pairing official status feeds with real-time user reports, so a spike in unofficial reports can trigger an alert before the provider posts anything public.
StatusGator detected a possible outage before Google Cloud’s official acknowledgment, then tracked the incident through to resolution.
Zonal outages are a classic case for crowd-sourced detection. Because the failure was confined to one zone, it did not look like a broad, platform-wide event, the kind that lights up a global status dashboard quickly. The numbers tell the story: Google later dated the start of impact to 06:08 UTC, the first user report reached StatusGator just 14 minutes later at 06:22 UTC, and StatusGator sent its Early Warning Signal at 07:02 UTC. Google’s own public acknowledgment did not appear until 08:37 UTC, roughly two and a half hours after the incident began. Subscribers had more than an hour and a half on the official notice to recognize the issue was Google’s, check their europe-west2 dependencies, and plan around it.
For historical context, you can review Google Cloud’s outage history and the Google Cloud outage map to see how often these events happen and where they land.
The blast radius: 171 downstream services affected
One power event in a single Google Cloud zone did not stop at Google. Because StatusGator monitors thousands of services, it watched the outage ripple outward in real time, with 171 services it tracks showing signs of impact during the same window. They fall into three groups, by how directly the evidence tied them to this outage. Each name links to its StatusGator page.
Confirmed affected (50): cited Google Cloud directly
These services named Google Cloud in their own status communications during the incident, the clearest evidence of impact.
- Cube Cloud
- Hostinger
- Purple
- PPRO
- Traders Connect
- Aiven
- Northflank
- Patients Know Best
- Elastic Cloud
- EMQ Technologies
- Nebula
- Camunda
- ThriftCart
- Spectra Networks
- MoonPay
- Go Cloud Communications
- JFrog
- Solace
- Statscore
- Flywheel
- MongoDB Cloud
- Collibra
- MongoDB
- Neo4j Aura
- Palo Alto Networks
- ClickHouse
- Sourcegraph
- Cockroach Labs
- Confluent
- Instaclustr
- Ravelin
- Starburst Galaxy
- Xano
- dbt Cloud
- Bringg
- Pentos
- OneID
- Acquired
- Intruder
- Genio
- TingleFlow
- SonicWall Cloud Secure Edge
- SonicWall
- DataStax Astra
- Person Centred Software
- StatusCake
- Integrated Finance
- Grafana
- Heron Data
- BlazeMeter
Likely affected (61): pointed to an upstream provider
These services reported problems tied to an upstream provider or the affected region during the same window, strongly consistent with this outage.
- Revolut
- Kagi
- Tink France
- Tink Germany
- OpenAthens
- Tink Norway
- Kinsta
- Sequence
- 2Chat
- Tink Italy
- Tink Spain
- EveryPay
- Qonversion
- iDEAL
- Bloomreach
- Akeneo
- Worldpay
- Tink UK
- Web Security Service
- Broadcom Cloud Secure Web Gateway
- Aiia
- Broadcom VIP
- Tink Netherlands
- Stripe
- eTailPet
- Mollie
- Broadcom Web Isolation
- Snowflake
- Maksekeskus
- CloudAMQP
- Conferma
- Cloudways
- Tink Portugal
- Pegasystems
- Tink XS2A
- Viva Wallet
- Yapily Platform
- Tink Denmark
- Paybis
- Tink Finland
- Prisma SASE
- Salt Edge
- Ledger
- Curve
- Xactus
- ThousandEyes
- Contentful
- Yapily Institutions
- PaperCut
- Blacksmith
- MultiSafepay
- PAYTEF
- Thomson Reuters Legal
- GoPay
- Tink Belgium
- SafeNet Trusted Access EU/US
- Solidgate
- Primer
- RunPod
- Hologram
- FTAPI
Possibly affected (60): a spike in user reports during the outage
These services saw a notable spike in user reports, or other generic signals, during the outage window, suggesting a likely connection to the same event.
- MyBib
- Google Antigravity
- Naukri
- Xendit
- Lloyds Bank
- PolyBuzz
- Pulsz
- Oanda
- Onfido
- Clark University – ITS
- Luno
- 3CX
- DPD
- Tazapay
- VirusTotal
- Arista CloudVision
- 1310.io
- DPD Local UK
- Opti9
- OnePipe
- Vinted
- DPD Tracking
- Elite Dangerous
- OVHcloud Web Hosting
- FanDuel
- WP Engine
- Catchpoint Systems Node Page
- Trello
- Zoho
- Sendcloud
- Bank of Scotland
- Spinwheel
- Krystal
- HUB2
- Cato Networks
- Broadcom Email Security
- Dye & Durham Unity
- Avigilon
- PayFit
- Royal Mail
- LINK Mobility
- Stax
- G-core
- Broadcom WebFilter
- Sky UK
- Kaseya
- ResDiary
- Hospitality
- Mono
- Shippeo
- Aitronos
- Spotify
- Plum Fintech
- Microsoft Copilot (Microsoft 365)
- Flutterwave
- Smobilpay
- NatWest
- Microsoft 365 Copilot Chat
- Sage
- Edgenuity
Lessons for teams that depend on Google Cloud
- Design for zone failure. A single zone losing power took down every service running in it. If your workload lives in one zone, a power event is a total outage. Spread critical workloads across zones or regions.
- Know your region and zone. A zonal outage in europe-west2-b does not show up as a global failure, but it takes down everything you host there. Monitor at the zone and region level.
- Do not wait for the official notice. Google dated the start of impact to 06:08 UTC but did not publicly acknowledge it until 08:37 UTC. Real user reports surfaced the problem far sooner.
- Turn on proactive alerting. Early Warning Signals exist for the window between a spike in reports and the official acknowledgment.
Frequently asked questions
Was Google Cloud down on October 20, 2026? Yes, in the europe-west2-b (London) zone. A power event took down dozens of services, including Cloud SQL, BigQuery, and Compute Engine, with user reports from about 06:22 to 10:07 UTC. For live status, check the Google Cloud status page on StatusGator.
What caused the Google Cloud outage? Google attributed it to a power event in its europe-west2-b zone. When the zone lost power, the services running in it went down together, which is why so many unrelated products failed at once.
Which Google Cloud region was affected? The outage was centered on the europe-west2-b zone in London, with wider elevated errors across the europe-west2 region. That geographic focus is why most reports came from the UK.
How long did the Google Cloud outage last? User reports spanned about three hours and 45 minutes, from 06:22 to 10:07 UTC, with Google posting its resolution later in the day.
How can I get warned about Google Cloud outages early? Monitor Google Cloud with StatusGator and enable Early Warning Signals, which alert you when user reports spike ahead of the provider’s public acknowledgment.
Monitor Google Cloud with StatusGator
If your workloads run on Google Cloud, a zonal or regional outage can take down your databases and apps while the rest of the platform looks healthy. StatusGator combines official status pages with real user reports and sends Early Warning Signals the moment a pattern emerges, often well before the provider posts anything public.
Start monitoring with StatusGator and get early warnings for the services your team depends on.
