Register / Data centre / EIR-1986
← Back to the list
Data centreUnited StatesEIR-1986

Google Cloud data centre UPS battery failure causes six-hour Ohio outage

Mar 2025Published 4 Oct 2026Updated 13 Oct 2026Responsible editor: Chief Editor

Estimated damageEstimated by the register's adjusterHow? Know the real figure?
Figure hidden: shown to subscribersLow confidence
out of service for about •• daysShown in the record →Full estimate
From the recordRoot cause hiddenFailed part hidden

Companies named in the reporting of this entry (being named is not a finding of fault): Google Cloud

On 29 March 2025 the us-east5-c zone of Google Cloud's US East region, centred on Columbus, Ohio, lost utility power. The uninterruptible power supply (UPS) system that should have bridged the gap until generators started suffered a critical battery failure and could not carry the load. Virtual machine instances in the zone lost power and went offline. The power outage and UPS failure also caused packet loss within the zone, affecting network communication, and made a limited number of storage disks unavailable.

Engineers were alerted at 12:54 Pacific time, diverted traffic away from the affected location where possible, and bypassed the failed UPS to restore power from a generator by 14:49. Most services recovered shortly afterwards, but some needed manual action for full recovery. The disruption lasted six hours and 10 minutes and affected more than 20 Google Cloud products, including Compute Engine, Cloud SQL and Persistent Disk.

Google's incident report attributes the outage to a loss of utility power, which triggered a cascading failure within the UPS system, followed by a critical battery failure. The Register reports that the UPSes appear to have prevented generator power reaching the racks until they were bypassed. Google has not publicly identified the cause of the battery failure and says it is working with its UPS vendor, which it did not name.

  1. Test UPS battery systems under realistic load, including the transfer to generator, because the battery is the link the design depends on during a utility failure.
  2. Ensure a failed UPS can be bypassed quickly and safely to generator supply, and rehearse that procedure so it does not depend on manual improvisation.
  3. Test failover of workloads to other zones and regions, since some services in this incident needed manual steps to recover.
  4. Map secondary dependencies such as storage, networking and control planes within a zone, because the effects of a power loss spread well beyond the power system itself.

More lessons on outages, backup supply and restoration, worker safety and safe systems of work, spacing, layout and fire spread, from every incident on the register.

Site pages: Google, Cloud Storage Corp (status, dates and other incidents)

Other entries that name Cloud

Published 2026-10-04. Written from public reporting. Descriptive, not investigative. See the data accuracy disclaimer.

See the register's stories first in Google: add it as a preferred source.

Chart to reuse with this story: Data centre incidents in the United States, by year, free with a credit (PNG).

Data centre incidents in the United States: 49 entries, with a free alert for the next one.

See UPS and battery room fires at data centres and backup generator failures at data centres.

Browse the full register or submit an incident.