What happened
A power fault at a Google Cloud facility in the Netherlands triggered a nearly 15-hour outage on July 15, 2026, taking down Google Cloud VMware Engine, Bare Metal Solution and Google Cloud NetApp Volumes in the europe-west4-a zone. Google's own Service Health incident report puts the disruption at 14 hours and 55 minutes, beginning at 16:39 PST on July 15 and clearing at 07:34 PST the following day.
Root cause
According to Google's incident report and reporting from Data Center Dynamics, the trigger was an electrical fault on the utility grid feeding the facility. The disruption interrupted power distribution and disabled cooling infrastructure. With cooling offline, temperatures inside the affected data halls rose quickly enough that Google shut down host servers, storage clusters and network switches as a protective measure rather than risk hardware damage from overheating.
Who was affected
Google's incident report puts the customer impact at 24 private clouds belonging to 20 distinct VMware Engine customers, plus 9 separate Bare Metal Solution customers, all within europe-west4. That is a narrower blast radius than a regional Compute Engine or networking outage would produce, since VMware Engine and Bare Metal Solution are specialized services used by enterprises running existing VMware or bare-metal workloads rather than by general Compute Engine tenants.
Why it matters
The outage is a reminder that even hyperscale cloud regions still depend on physical infrastructure with real failure points: a single utility-side electrical event was enough to take a zone offline for the better part of a day. The Register's coverage of the incident makes a related point worth repeating for anyone planning disaster recovery around a single cloud zone: published SLAs and marketing language around "always-on" infrastructure describe target availability, not a guarantee that a single physical facility can't go dark. Teams running VMware Engine or Bare Metal Solution workloads in a single zone without a documented failover to a second zone or region were exposed for the full outage window with no cloud-side mitigation available to them.
What to check
Anyone running production workloads on Google Cloud VMware Engine, Bare Metal Solution or NetApp Volumes in europe-west4 should pull Google's post-incident report directly rather than rely on summary coverage. Teams should confirm whether their own architecture has cross-zone or cross-region failover configured for these specific services, since not all of Google Cloud's specialized infrastructure products support the same multi-zone redundancy patterns as core Compute Engine, and should check contractual SLA credit eligibility if the outage caused a measurable service disruption.
Sources
- Google Cloud Status, incident 3BvH3LVGcupoYqV6F4Nw - https://status.cloud.google.com/incidents/3BvH3LVGcupoYqV6F4Nw
- Data Center Dynamics, "Google Cloud experiences outage after power and cooling failure in Netherlands data center" - https://www.datacenterdynamics.com/en/news/google-cloud-experiences-outage-after-power-and-cooling-failure-in-netherlands-data-center/
- The Register, "Google Cloud outage shows it's still hard to understand hyperscalers' real resilience regimes" - https://www.theregister.com/off-prem/2026/07/21/google-cloud-outage-shows-its-still-hard-to-understand-hyperscalers-real-resilience-regimes/5275405
- Computing.co.uk, "Google Cloud outage exposes datacentre resilience concerns" - https://www.computing.co.uk/news/2026/datacentre/google-cloud-outage-exposes-datacentre-resilience-concerns