Site Hashrate & Uptime (Avg ÷ Spec) — trailing 7 / 30 / 90 days
Regenerated 2026-08-25, ~12:00 UTC. Freshest telemetry point in this pull: 2026-08-25 11:50 UTC (InfluxDB
MARAGON_Hashrate). 13 sites, 46 pods, roster verified against master_config.json. Sorted by
7-day uptime, highest first.
"Uptime" here = period average hashrate ÷ spec (installed) hashrate — a utilization ratio, not
the binary threshold measure in yesterday's site-uptime-summary-2026-08-24.html. A site can be up 100%
of the time and still show a low ratio here (auto-sleep, curtailment, thermal derate) — low numbers aren't
automatically a fault. Fleet cols (right, new): Spec Contrib% = this site's share of total installed
fleet spec (2,350,620 TH/s across all 13 sites); Fleet Uptime Cont% = that share realized at the site's 7-day uptime
rate (Spec Contrib% × 7d Uptime%); Fleet Uptime Miss% = Spec Contrib% − Fleet Uptime Cont%, i.e. the
slice of total fleet capacity this site is currently leaving on the table. The three sum down to 100% / 59.5% /
40.5% — that 59.5% is the fleet-wide capacity-weighted 7-day uptime.
Spec = installed miners (inv_miner_count, fallback design miner_count) × 180 TH/s (M60 spec). Osprey/West TX TBD scored 0% uptime for the fleet columns (see notes).
| Site |
Pods |
Spec
(TH/s) |
7d Avg
(TH/s) |
7d
Uptime% |
30d
Uptime% |
90d
Uptime% |
Spec
Contrib% |
Fleet
Uptime Cont% |
Fleet
Uptime Miss% |
Method & definitions
Spec (TH/s): installed miner count per pod (inv_miner_count, fallback design
miner_count for John, Dan, West TX TBD) × 180 TH/s, summed per site.
Avg hashrate / Uptime%: InfluxDB MARAGON_Hashrate, site's pods summed per 15-minute
window, averaged over all windows in the period with at least one reporting pod (sleeping miners count as 0, not
excluded). Uptime% = that average ÷ Spec. One 90-day pull per pod, sliced into the three trailing
windows.
Fleet columns: Spec Contrib% = site Spec ÷ 2,350,620 (sum of all 13 sites' Spec). Fleet
Uptime Cont% = Spec Contrib% × 7d Uptime% ÷ 100. Fleet Uptime Miss% = Spec Contrib% − Fleet
Uptime Cont%. Ranks sites by how much total-fleet capacity each is actually losing, not just by its own percentage —
e.g. GN's 64.8% uptime is mid-pack, but at 25.6% of fleet spec it's the single largest miss (9.0 fleet points);
Osprey and West TX TBD are non-producing buildout sites with no live uptime figure, so both are scored 0% here and
their full spec share (2.7% and 8.7%) shows as 100% fleet miss.
Flags & data-quality notes
Stout 90d: Stout 3/4 (480 miners each) only start reporting 2026-07-17 (~39 of 90 days). The 90-day
figure blends a 2-pod period against a 4-pod spec, mechanically dragging it to 46.1% vs. 74–75% at 7d/30d —
capacity growth, not a performance decline.
WW†: Influx history starts 2026-07-30 (~26 days old). 30d/90d figures are averaged only over
the days WW has existed — its actual average since coming online, not a true 90-day average.
Ellyson: spec uses the 84 miners actually installed on Ellyson 2, not its 185-miner design
figure.
Osprey: measured, genuine 0% across all windows — non-producing buildout site, not a gap.
West TX TBD: zero telemetry anywhere in the 90-day window (dark since 2026-05-07) — shown n/a, not
0%, on the per-site columns; scored 0% only in the fleet columns since it contributes nothing live today.