InfoWorld published a column by David Linthicum on 23 October about a recent Google Cloud outage in London. Google said a power problem hit its data centers in the europe-west2 region. Linthicum wrote that it began early in the day. In his account, banks, delivery firms, e-bike providers, budgeting apps, some NHS app appointment services and court video links were all caught in it. Google later said most affected workloads had recovered, with only a few customers still hit.
The column gives no date and no length for the outage. It says nothing on the cause beyond Google's description, and it gives no count of affected customers. It also does not say which listed services ran on Google Cloud directly and which reached it through other platforms. Linthicum says both kinds of dependence were in play.
WebPulse's reading: the risk is the chain of hidden dependencies, not the provider's uptime. Linthicum says services that look unrelated can share a cloud provider, a region, identity systems and APIs. He suggests asking which suppliers share your provider or region, which of their services are regional or global, and which workloads can run in a reduced mode. He adds that the company, not the provider, owns its failover and must test it.