# #2783 firstmin public-RPC 429s (17 Oct 02:25 BKK)
Evidence (/opt/fm_mirror): all 27 lines are ONE burst 17 Oct 02:04:07-02:04:58 UTC (mc5, wl_wlink, wl_wlink1 at once; feed_gap 12-16 s). No other 429 lines that hour.
Same minute, research-box v4tape/pretag/lpowntape and main aipick/tracker also got 429 -> the shared public node throttled everyone; not one noisy caller.
Underlying cause: the dRPC first-RPC answers HTTP 403 (billing/cap) every minute since 16 Oct 00:26 UTC on wlink/wlink1 (mc5 since 8 Oct). So all 3 live processes run on the public node only, about 150-175 calls/min each, with no second source. Normal load is fine (failed calls 1-3/min), but a provider hiccup has no fallback, so gaps appear.
Fix at the cause = restore the dRPC key (billing/cap) or give live a second free provider. Both are live config, exec/Ben only. Nothing changed by me.