Uptime history

Each bar represents one calendar day in UTC. A green bar means every probe in that 24-hour window completed inside the operational thresholds. A yellow bar means one or more probe windows reported elevated latency or transient failures, but the mirror remained reachable. A red bar means descriptor unreachability for at least one probe window, and the mirror was effectively offline during part of that day.

A red bar does not mean the market itself was down. With three independent mirrors, a single outage on one onion is absorbed by the other two without any user-facing impact. The bars track per-mirror reachability, which matters for capacity planning and operator post-mortems, not for end-user availability.

The probe network operates from inside Tor with no clearnet leg. Probe results are signed by an observer key, witnessed quarterly by independent custodians, and republished on each mirror under /announcements/probes for anyone who wants to verify the raw data themselves.

We measure descriptor reachability, login page latency, and signature freshness. Those three together approximate the experience of a user opening a mirror in Tor Browser. We do not measure escrow contract throughput, vendor inventory updates, or buyer-side wallet sync. Those happen inside the market core and are tracked separately by the platform team.

We do not measure clearnet anything. There is no clearnet path to Nexus Market. Anyone offering a clearnet status page for the market is either confused or malicious. The only legitimate surface is this page, signed under 0A9D.

Probes share no identity with user sessions. They use throwaway circuits, never log in, never load product pages, never touch the order subsystem. The probe footprint is minimal and uncorrelated with user activity, which is the right design for a public health page that should not introduce any new attack surface to the market itself.