The teams and operators behind these public interfaces.
138 matching of 138 loaded providers · verifying sources · registry
One operator often runs surfaces for several subnets, and one subnet often depends on several operators. Modelling them separately is what makes it possible to see that an outage spans four subnets because they share a host, rather than reading it as four unrelated incidents.
A provider is attached to a surface only where the operator declares it or a first-party source shows it. Where ownership is unclear the surface stays unattributed rather than being assigned on a guess, because a wrong attribution is a claim about somebody's business.
Uptime here is the share of our probes that got a usable answer over the window, not an operator's SLA. A provider running one endpoint that has never failed and a provider running forty with one flapping host can show similar percentages, so the endpoint count beside it is part of the reading.
https://api.metagraph.sh/api/v1/source-health