01

Location of the observer

A probe region identifies where our measurement originates. It is not a statement about the provider data center handling the request. Providers can route requests internally in ways this monitor cannot see. Seoul describes a planned observation location, not a claim that a model runs in Seoul.

02

Your path can differ

A person on a mobile connection in the same city may take a different route from a cloud probe. Browser behavior, connection reuse, network conditions and product choice affect the experience. Regional observations provide context, but are not a personalized diagnosis of a home connection or internet service provider.

03

Coverage has to be earned

A region stays planned until it has an operating probe and real records. It should not inherit a green status from another region. Missing regions stay missing in coverage statements. A global label would be especially misleading when observations exist in just one location.

04

A useful comparison

Compare the same model and experiment settings over overlapping windows. Show sampling imbalances between locations. Region-specific history must support region-specific recommendations. During this release all regional pages explain the plan and limitations; they do not display invented values or claim that one network is faster.

05

Sources and further reading

Our methodology · More guides