dangernoodle-io / TaipanMiner / 24907617969

90%

push

github

feat(asic): telemetry-drop diagnostics + ui health state (TA-223) (#180)

chip 0 on bitaxe-650-1 trips the ta-221 sanity clamp every 5s poll
(6 warn lines/cycle, 72/min). the clamp prevents buffer poisoning
but the current log emits only the computed ghs, not the inputs to
the math — can't tell uart-corruption / silicon-defect / routing-
misfire apart without more data.

firmware:
- 30s cooldown per (chip, metric) on the sanity-fail warnings —
  stops the log spam
- log format now includes delta (hex), elapsed seconds, and raw
  asic_addr byte (rx[6]) so the routing-misfire hypothesis is
  testable from a serial capture
- per-chip drop counters (total_drops, error_drops, domain_drops[4])
  added to s_chip_meas, exposed on /api/stats.asic_chips[] payload
- paired total/error drop policy: if either metric fails the clamp,
  drop both so the 1m/10m rolling buffers stay paired

webui:
- ChipsCard surfaces a warning state (var(--danger)) when
  total_drops > 0, labeled "telem corrupt (N drops)" with tooltip
  explaining the chip is still hashing. previously the card dimmed
  chip 0 as if it were inactive, conflating "dead" with "hashing
  but reports garbage"

diagnostic only — does not fix the underlying chip-0 corruption;
this PR captures the data we need to decide the fix.

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

197 of 252 branches covered (78.17%)

Branch coverage included in aggregate %.

656 of 673 relevant lines covered (97.47%)

617895.43 hits per line