Mostly operational

Uptime is the share of probes that succeeded in the last 24 hours; the median is the probe’s own round trip, not your request’s. When a provider fails here, routing chains fall back to the next model automatically - an incident on one provider is a slowdown on Final Router, not an outage. During high demand a provider may answer more slowly than usual; requests keep succeeding, falling back where a chain allows it.