slotwatch · compare

how fast we read chain state vs other public sources

All probes fire in parallel from the slotwatch process on Ashburn, each through the same fresh HTTP client so connection-pool warmth is uniform — apples to apples. The two self HTTP rows hit 127.0.0.1:8053 over loopback: /health shows pure axum overhead (no upstream call), /current shows what a consumer asking "who's leading now?" pays end-to-end. The external rows make a live getSlot JSON-RPC call. Reference slot is whatever /current just returned, and slot lag = source − reference (negative = behind).

Current Snapshot

probing…
Source Region Kind Slot Lag RTT Status
first probe in flight…

Latency (this probe)

round-trip ms from Ashburn

Rolling Stats (last 30 probes)

building history…
Source n p50 p95 p99 min max fail %
collecting samples…