The health page at oxyon.io/xl1/health (JSON: https://oxyon.io/api/xl1/{mainnet|sequence}/health) gives two answers, kept apart:
- The XL1 network, as observed: how the chain itself is doing.
- Oxyon's own observers and telemetry: whether Oxyon's instruments are healthy.
A stalled Oxyon observer does not make the network look stalled, and a slow chain does not look like an Oxyon fault.
Checks and fixed thresholds
| Group | Check | OK | Slow | Stalled |
|---|---|---|---|---|
| network | Newest block, time since | ≤ 120 s | ≤ 600 s | > 600 s |
| network | Block time, median of the last 100 | ≤ 30 s | ≤ 120 s | > 120 s |
| network | Active producers (1 h), busiest share | information | ||
| network | Racing, never winning | information | ||
| oxyon | Race observer, its own published state | observing | stale (not rebuilt for 30 min) or unverified | stalled (analysis 10+ min behind) or no state |
| oxyon | Observer self-recovery | information | reports that it needs a person | |
| oxyon | Oxyon producer telemetry (self-reported), with each report's age | information; a report older than 30 min is flagged stale |
- Boundaries are inclusive. 120 s is OK; 121 s is slow.
- "Unknown" means a source couldn't be read. It is never shown as OK.
- Producer telemetry is information only. A stale self-report is flagged, but it never makes the network look unhealthy: it says something about Oxyon's own reporting, not about XL1. A report whose time can't be read is shown as unknown, never as fresh.
- Each group's grade is the worst graded check in it. Information rows never change a grade.
Where the answer comes from
- Only from data the explorer already has: Oxyon's chain watcher and the race observers' published state.
- Opening the page contacts no block producer or observer. The answer is cached for 15 seconds.
Not checked
XYO RPC reachability isn't checked here: this view makes no RPC call. Pages that use the RPC, such as an address's balance and transfers, report their own errors when it fails.