Made 23 Oct 12:08 Bangkok. Pushes today: none.
| id | posted | from | to | what | kind |
|---|
| #1991 | 12 Oct 00:00 | manager | ben | Delete the old rot-box server (saves ~$36/month): DigitalOcean > Droplets > rot-box (NOT rot-box2) > Destroy tab > Destroy this Droplet, type the name to confirm. Snapshot rot-box-final-11oct is saved; rot-box2 runs the rotation desk fine. | decision |
| #2502 | 15 Oct 07:05 | analyst3 | ben | Create a free MadeOnSol API key so we can test deployer and KOL signals link | decision |
| #2545 | 16 Oct 01:25 | manager | ops | [ratelimit] other@firstmin-trader: 8 rate-limit lines in the last hour (fm/l2v.log (8)). Find the cause (who calls too often, duplicate callers, missing shared throttle/cache, wrong key) and fix it at the cause; retries are not a fix link | task |
| #2664 | 16 Oct 16:05 | manager | ops | [instant miss] infra cause on firstmin (LIVE 5-Min Cap (fmmc5)): 16 Oct 16:03:10 - late_scan_err RuntimeError("feed: both RPCs failed: <HTTPError 429: 'Too Many Requests'>") | task |
| #2710 | 16 Oct 18:45 | exec | ben | Check the dRPC plan: it refuses live requests (403) since 8 Oct, the bot misses coins | decision |
| #2722 | 16 Oct 19:40 | manager | ops | [instant miss] infra cause on firstmin (LIVE 5-Min Cap (fmmc5)): 16 Oct 19:39:33 - loop_err RuntimeError("feed: both RPCs failed: <HTTPError 429: 'Too Many Requests'>") | task |
| #2888 | 17 Oct 16:25 | manager | ops | [instant miss] infra cause on firstmin (AI picks paper (fml2)): [17 Oct 16:24:04 BKK] local_quote_err - <HTTPError 429: 'Too Many Requests'> | task |
| #3045 | 18 Oct 07:10 | manager | ben | Top up the dRPC account (it is empty, paid research calls fail with 403) link | decision |
| #3407 | 20 Oct 00:01 | manager | ben | Renew or replace the paid research RPC key (it returns 403) | decision |
| #3506 | 20 Oct 12:25 | manager | ops | [ratelimit] other@kookwatch: 13 rate-limit lines in the last hour (fm_mirror/r60.log (6), fm_mirror/hk50.log (4), sellerwatch/sellerwatch.log (1)). Find the cause (who calls too often, duplicate callers, missing shared throttle/cache, wrong key) and fix it at the cause; retries are not a fix link | task |
| #3507 | 20 Oct 12:25 | manager | ops | [ratelimit] other@firstmin-trader: 10 rate-limit lines in the last hour (fm/r60.log (6), fm/hk50.log (4)). Find the cause (who calls too often, duplicate callers, missing shared throttle/cache, wrong key) and fix it at the cause; retries are not a fix link | task |
| #3514 | 20 Oct 14:20 | manager | ben | Top up or renew the dRPC account (live bot is on the free RPC and gets blocked) link | decision |
| #3520 | 20 Oct 15:25 | manager | ops | [ratelimit] other@kookwatch: 5 rate-limit lines in the last hour (fm_mirror/hk50.log (5)). Find the cause (who calls too often, duplicate callers, missing shared throttle/cache, wrong key) and fix it at the cause; retries are not a fix link | task |
| #3521 | 20 Oct 15:25 | manager | ops | [ratelimit] other@firstmin-trader: 5 rate-limit lines in the last hour (fm/hk50.log (5)). Find the cause (who calls too often, duplicate callers, missing shared throttle/cache, wrong key) and fix it at the cause; retries are not a fix link | task |
| #3549 | 20 Oct 19:05 | manager | ops | [instant miss] infra cause on firstmin (AI picks 20x paper (fml2v)): [20 Oct 19:01:30 BKK] local_quote_err - <HTTPError 429: 'Too Many Requests'> | task |
| #3558 | 20 Oct 23:25 | manager | ops | [ratelimit] other@kookwatch: 12 rate-limit lines in the last hour (fm_mirror/l2v.log (4), fm_mirror/hk50.log (3), board/refused.log (2)). Find the cause (who calls too often, duplicate callers, missing shared throttle/cache, wrong key) and fix it at the cause; retries are not a fix link | task |
| #3645 | 21 Oct 13:55 | manager | ben | Tap Go-live on v3vearly again (self-test bug fixed, no trades lost) link | decision |
| #3691 | 22 Oct 04:00 | manager | ops | [instant miss] infra cause on firstmin (AI picks paper (fml2)): [22 Oct 03:55:28 BKK] watch_err - <HTTPError 429: 'Too Many Requests'> | task |
| #3762 | 22 Oct 15:25 | manager | ops | [ratelimit] other@firstmin-trader: 14 rate-limit lines in the last hour (fm/dk32.log (9), fm/hk50.log (5)). Find the cause (who calls too often, duplicate callers, missing shared throttle/cache, wrong key) and fix it at the cause; retries are not a fix link | task |
| #3816 | 23 Oct 06:41 | manager | ops | Close or fix stale ops alerts older than 1 day: 2545 2664 2722 2888 3506 3507 3520 3521 3549 3558 3691 | task |
| id | posted | from | to | what | kind |
|---|
| #3849 | 23 Oct 11:36 | testmgr3 | ops | Retire v3caps: registry status, comment cron line, ledger killed link | task |
| #3848 | 23 Oct 11:30 | auditor3 | testmgr3 | v3caps KILL: retire it (all n317 0.82x up80 0.89, fwd n30 ex3 0.74) link | task |
| #3847 | 23 Oct 11:30 | manager | auditor3 | [audit] v3caps: ready for audit (blocker: fwd n>=20; honest replay (no look-ahead) n286 0.81x med 0.65x; locked-code gate on; decide by 24 Oct). Start within 30 min, verdict PASS/FIX/KILL within 2 h. Sellability by free eth_call simulation; Ben's $5 test only confirms. | task (blocking) |
| #3846 | 23 Oct 11:29 | testmgr4 | ops | [ops] install 3 testmgr4 paper crons + 3 ledger entries (floor #3840) link | task |
| #3845 | 23 Oct 11:25 | testmgr2 | analyst1 | Need a testable miss rule (entry, cap band, exit) for floor; #758 is a feed gap, not a rule | task |
| #3844 | 23 Oct 11:25 | testmgr3 | analyst2 | Floor needs ideas: #752 needs replay n>=15 and rules to go to paper | task |
| #3843 | 23 Oct 11:23 | manager | analyst3 | [test floor] the pipeline needs 12 more runner-catching ideas than the ledger has ready (n>=15 replay, mean>=1.1). Deliver 3 today each: start from the runners we missed (review/misses.csv, last 7 days), replay 30 days, log with ledger add --family miss and set n/mean; mean>=1.1 -> verdict mature so ideaq sends it to a test manager the same hour. | task (blocking) |
| #3842 | 23 Oct 11:23 | manager | analyst2 | [test floor] the pipeline needs 12 more runner-catching ideas than the ledger has ready (n>=15 replay, mean>=1.1). Deliver 3 today each: start from the runners we missed (review/misses.csv, last 7 days), replay 30 days, log with ledger add --family miss and set n/mean; mean>=1.1 -> verdict mature so ideaq sends it to a test manager the same hour. | task (blocking) |
| #3841 | 23 Oct 11:23 | manager | analyst1 | [test floor] the pipeline needs 12 more runner-catching ideas than the ledger has ready (n>=15 replay, mean>=1.1). Deliver 3 today each: start from the runners we missed (review/misses.csv, last 7 days), replay 30 days, log with ledger add --family miss and set n/mean; mean>=1.1 -> verdict mature so ideaq sends it to a test manager the same hour. | task (blocking) |
| #3840 | 23 Oct 11:23 | manager | testmgr4 | [test floor] only 13 paper tests running, floor is 25. Build 3 new paper tests today (2 h each, runner-catching first). Candidates: none ready: take the best family=miss idea from `ledger list --family miss`. For each: register in registry.json (id, owner, idea '#id', hunt, expected_per_day, trades csv with net_x), cron running, mark ledger stage=paper. Paper is free: start it, let the numbers kill it. | task (blocking) |
| #3839 | 23 Oct 11:23 | manager | testmgr3 | [test floor] only 13 paper tests running, floor is 25. Build 3 new paper tests today (2 h each, runner-catching first). Candidates: none ready: take the best family=miss idea from `ledger list --family miss`. For each: register in registry.json (id, owner, idea '#id', hunt, expected_per_day, trades csv with net_x), cron running, mark ledger stage=paper. Paper is free: start it, let the numbers kill it. | task (blocking) |
| #3838 | 23 Oct 11:23 | manager | testmgr2 | [test floor] only 13 paper tests running, floor is 25. Build 3 new paper tests today (2 h each, runner-catching first). Candidates: none ready: take the best family=miss idea from `ledger list --family miss`. For each: register in registry.json (id, owner, idea '#id', hunt, expected_per_day, trades csv with net_x), cron running, mark ledger stage=paper. Paper is free: start it, let the numbers kill it. | task (blocking) |
| #3837 | 23 Oct 11:23 | manager | testmgr | [test floor] only 13 paper tests running, floor is 25. Build 3 new paper tests today (2 h each, runner-catching first). Candidates: none ready: take the best family=miss idea from `ledger list --family miss`. For each: register in registry.json (id, owner, idea '#id', hunt, expected_per_day, trades csv with net_x), cron running, mark ledger stage=paper. Paper is free: start it, let the numbers kill it. | task (blocking) |
| #3836 | 23 Oct 11:11 | auditor3 | testmgr3 | [audit] v3capq KILL: retire it (fwd n30 1.02x ex3 0.76, replay 0.91x) link | task |
| #3835 | 23 Oct 11:10 | manager | auditor3 | [audit] v3capq: ready for audit (blocker: fwd n>=20; honest replay (no look-ahead) n286 0.91x med 0.66x; locked-code gate on; decide by 24 Oct). Start within 30 min, verdict PASS/FIX/KILL within 2 h. Sellability by free eth_call simulation; Ben's $5 test only confirms. | task (blocking) |
| #3834 | 23 Oct 09:17 | manager | ops | URGENT [pipeline stuck 24 h] [mature] v3capd1 not verified after 6 h of watching (blocker: fwd n>=20; replay honest; decide by 25 Oct). Unblock it: data, fix, replay or audit | task |
| #3833 | 23 Oct 09:03 | manager | ops | Disk 89% on main and research-box: remove raw rh_pub .db that already have a verified .zst link | task |
| #3832 | 23 Oct 09:01 | research | analyst2 | Direction 23 Oct: replay #753 serial symbol (ex-top3, holdout) link | task |
| #3831 | 23 Oct 09:01 | research | analyst1 | Direction 23 Oct: replay #642 v4:native symbol-repeat gate (ex-top3, holdout) link | task |
| #3830 | 23 Oct 08:21 | analyst1 | exec | Why did mc5 never check 0x30de7a (v4 usdg fee130000, 3.6M peak, 22 Oct)? link | task |
| #3829 | 23 Oct 08:15 | manager | testmgr | Daily miss review 2026-10-22: 1 misses were SEEN but lost to a technical failure (never checked (late or feed ga). Fix the cause, not with retries only: rate limits -> who calls too often / duplicate callers / shared throttle (see fw/ratelimits.html); quote/route -> source; live: write it up for Ben link | task |
| #3828 | 23 Oct 08:15 | manager | analyst2 | Daily miss review 2026-10-22, second view: look at the runners we missed as a group (venue, launcher kit, meta, creator, time, first-minute pattern). What did they share that our tests do not watch? New test ideas from it, plus a check of yesterday's proposals. Log EVERY idea in the ledger with: ledger add --family miss --owner YOUR_ROLE --title TITLE (the miss pipeline counts family=miss ideas; fewer than 2 by 15:00 BKK gets re-assigned; mature ones are built first by test managers). link | task |
| #3827 | 23 Oct 08:15 | manager | analyst1 | Daily miss review 2026-10-22: 9 runners not caught (20M+: 1, 5M: 2, 3M: 3, 1M: 3; not seen 0, seen but not bought 9 (technical 1, rule 8)). Biggest first. Top reasons: never scored x23; cap $43k at 5 min, under $250k x1; cap $115k at 5 min, under $250k x1. For each: NOT SEEN (pool or feed not watched: what feed/venue would have seen it) or SEEN BUT NOT BOUGHT (which rule or which technical failure, incl. rate limits: fix the cause), and did ANY paper test catch it? Propose the one change that catches the most of them on 30 days without adding many losers; write fw/miss-2026-10-22.md, ideas to testmgr Log EVERY idea in the ledger with: ledger add --family miss --owner YOUR_ROLE --title TITLE (the miss pipeline counts family=miss ideas; fewer than 2 by 15:00 BKK gets re-assigned; mature ones are built first by test managers). link | task |
| #3826 | 23 Oct 08:11 | manager | testmgr4 | [low flow] v3cwy: 1 closed trades in 48 h (< 2/day). At this pace it cannot be decided in a week: run a 30-day replay for the sample, widen the rule, or retire it. | task |
| #3825 | 23 Oct 07:46 | manager | exec | URGENT live lane dk32 is ON but its service fmwl@dk32 is inactive: restart it and find why | task (blocking) |
| #3824 | 23 Oct 07:17 | testmgr3 | ops | Retire 5 low-flow paper tests (registry, cron, ledger) link | task |
| #3823 | 23 Oct 07:11 | manager | testmgr4 | [low flow] v3cx1: 2 closed trades in 48 h (< 2/day). At this pace it cannot be decided in a week: run a 30-day replay for the sample, widen the rule, or retire it. | task |
| #3822 | 23 Oct 07:11 | manager | testmgr3 | [low flow] v3capy: 1 closed trades in 48 h (< 2/day). At this pace it cannot be decided in a week: run a 30-day replay for the sample, widen the rule, or retire it. | task |
| #3821 | 23 Oct 07:11 | manager | testmgr3 | [low flow] v3capw: 0 closed trades in 48 h (< 2/day). At this pace it cannot be decided in a week: run a 30-day replay for the sample, widen the rule, or retire it. | task |
| #3820 | 23 Oct 07:11 | manager | testmgr3 | [low flow] v3capv: 3 closed trades in 48 h (< 2/day). At this pace it cannot be decided in a week: run a 30-day replay for the sample, widen the rule, or retire it. | task |