London cloud outage shows limits of relying on provider resilience
A power event in Google Cloud’s London region disrupted services used by banks, delivery providers, the NHS app, and court video systems. The outage illustrated how shared cloud infrastructure and dependencies can spread failures across seemingly unrelated organizations. Enterprises should design, fund, test, and own their recovery plans instead of assuming a hyperscaler automatically provides resilience.
A power problem in Google Cloud’s London region (europe-west2) disrupted several unrelated-looking services, including banks, delivery and e-bike providers, budgeting apps, NHS app appointment functions, and court video links. Google reported that most workloads recovered, though a few customers remained affected.
The event highlighted shared dependencies that can link organizations: managed services, networks, identity systems, APIs, and operational processes. Hyperscalers supply regions, zones, replication, backups, and monitoring, but these are components; application design, failover testing, and recovery ownership remain with the enterprise.
The outage could affect ordinary users less as direct Google customers than as patients, bank customers, delivery customers, court participants, and staff relying on third-party platforms. Repeated disruptions may erode confidence in digital public services and prompt organizations to fund tested continuity plans, though the wider social impact will depend on how often such shared dependencies fail and how quickly services recover.