Hour audit: 2026-09-03T21
back to the files for this hour · how the archive is audited
Every number on this page is a stored measurement read from this hour's own records: its
manifest.json, its witness
statistics sidecar, and the corpus attestation. Nothing here is computed from the served
bytes at the time you load the page, so no line below claims the bytes are correct. The
digest and the one command that checks them are printed instead.
No findings
Every check that could be evaluated for this hour passed. That is not the same as the hour being perfect: it means the records this archive keeps about it agree with each other and with the corpus ledger.
The report
POLYMARKET ORDERBOOK ARCHIVE — HOUR AUDIT
hour 2026-09-03T21 · corpus v3 · verdict AUDIT PASS
COVERAGE minutes present 60/60 · first event 2026-09-03T21:00:00.000051364+00:00 · last event 2026-09-03T21:59:59.999946525+00:00
VOLUME 73,141,554 rows · 735,761,486 bytes (701.68 MiB) · 7 products, all holding rows
WITNESSES bceh — 4 machines contributed rows
captured h: 54,486,783 · c: 53,183,354 · e: 46,606,298 · b: 41,050,433 (each machine's own feed; these overlap and do not sum to the file)
supplied h: 15,346,244 (20.98%) · c: 9,638,691 (13.18%) · e: 46,605,947 (63.72%) · b: 1,550,672 (2.12%) (rows of the merged file that came from it; these partition it)
3 product-machine pairs publish no supplied count; the figures above still sum to the file's 73,141,554 rows exactly, so those pairs supplied none
4 expected, all 4 contributed
INTEGRITY sha256 0c7d594fa512f5942ce42ab1832d88ceeee4ca8a52a93e45a3b4aacb45c407b7
the corpus checksum ledger names this file with the same digest
joined 73,142,057 to merged 73,141,554, minus 503 (minus 0.0007%)
boundary drops 503 · guard cache
SIDECARS none: this hour's manifest has no v31 key. The v3.1 sidecars are optional, and hours before them have none
MARKETS not measured (the per-hour market census runs on the capture host and its output is not published per hour)
HISTORY merged 2026-09-03T22:45:40Z as bceh
layout union-event-sorted-v1
re-merges not measured (the merge ledger records the merge that produced this file, not the history of attempts)
supersessions not measured (supersession is recorded per corpus in the storage registry, not in the hour manifest)
CHECKS 8/8 pass
PASS all products present
PASS row counts reconcile
PASS byte ranges tile the file
PASS the merge delta is explained
PASS every expected machine contributed
PASS the checksum ledger names this file
PASS the hour is whole
PASS a boundary guard ran
UPSTREAM an incident recorded by the venue intersects this hour: Venue paused trading, then twice more in cancel-only mode, 2026-09-03T17:06Z to 2026-09-03T21:35ZThe words in this report
Each line in the CHECKS block states what was found, not what was
hoped for: a check that passes is named for the thing that is true, and a check that does
not is renamed to the thing that is wrong. So FINDING never sits beside a
sentence asserting the opposite of the one below it.
The report above is fixed-width text, so nothing in it can be clicked. The same terms are linked here instead. Its verdict is a statement about minutes and nothing else. The row it reports comes from this hour's own manifest and its witness-statistics sidecar, and the completeness line comes from the corpus attestation. The gap between the joined and merged counts is deduplication, not loss, and the boundary drops are the duplicates discarded at the seam with the hour before. Where the history line names a supersession, this hour was rebuilt to a better state after it was first published. The precision is the timestamp resolution these files carry.
Detail
The same records, per product. A product holding no rows renders its zero rather than
being dropped: dropped and empty are indistinguishable to a reader, and only one of them is
a defect. A cell reading not meas. is one nothing published, which is a third
thing again and never a zero.
The class column is the merger's own verdict for that product, and it is
keyed to CONTRIBUTION rather than presence. FULL means every machine expected to
be capturing this hour delivered rows for that product; PARTIAL means some did
and some did not; SINGLE-BY-DESIGN means only one machine was ever expected, and
SINGLE-BY-DEFECT means several were and one delivered. So PARTIAL
here is a statement about the fleet, not about the product being damaged or truncated: the
rows that are present are whole.
PRODUCT rows bytes row groups sha256 best_bid_ask 9,855,430 109,384,701 0-37 3244c13805979d62ede02d134d11fcad520eb4eeb987b8d07ce888ade3593209 book 591,136 55,338,126 38-40 76703d83976f22bea2ee064998d9535c4e7400d0ce7e86327ed77c2c7a7e8dee last_trade_price 68,299 3,822,509 41-41 7592a6264c80d8d30652b3f0eaa3af4a74c3e558a3d8e718a7dd9c6fd3f2f806 market_resolved 1,697 242,001 42-42 46587db88d26d096422f921170b64abcb4f4827b6c918385b3633eecc2757a4d new_market 682 75,119 43-43 233d8e08f040042706df13aadaa9bd17d9173266518834998533594c536366b5 price_change 62,621,294 565,951,634 44-282 bad71d30cd96da9f36a55c00fc8344db04c25284238139e27cb343b565d4858c tick_size_change 3,016 109,617 283-283 9125e9b9477bc4e891d1b2d4db3b8afaee1ed3e6d302993c6abfaa723e9fe104 EVENT MIX count share price_change 62,621,294 85.62% best_bid_ask 9,855,430 13.47% book 591,136 0.81% last_trade_price 68,299 0.09% tick_size_change 3,016 <0.01% market_resolved 1,697 <0.01% new_market 682 <0.01% PER PRODUCT joined merged boundary exception mkts class best_bid_ask 9,855,530 9,855,430 100 3,635 FULL book 591,146 591,136 10 184,954 FULL last_trade_price 68,300 68,299 1 355 FULL market_resolved 1,697 1,697 0 38 FULL new_market 682 682 0 0 FULL price_change 62,621,686 62,621,294 392 109,899 FULL tick_size_change 3,016 3,016 0 76 FULL ARRIVAL SKEW m rows won p50 p90 p99 max unmeasurable interpretation best_bid_ask b 9,568,119 18 50ms 71ms 807ms 31.5s 0 transport_latency_assumed best_bid_ask c 8,799,626 692 4ms 13ms 59ms 6.1s 0 transport_latency_assumed best_bid_ask e 9,817,569 9,817,503 43ms 82ms 1.8s 33.0s 0 transport_latency_assumed best_bid_ask h 9,610,462 37,217 17ms 23ms 147ms 7.3s 0 transport_latency_assumed book b 145,873 11,740 157ms 1.2min 3.2d 78.5d 0 book_staleness book c 552,496 417,770 1.7min 8.2h 29.6d 97.0d 0 book_staleness book e 138,802 138,798 101ms 13.8s 7.6h 15.3d 0 book_staleness book h 146,498 22,828 37ms 1.5s 5.4h 97.0d 0 book_staleness last_trade_price b 61,006 8 99ms 5.3s 16.6s 31.4s 0 transport_latency_assumed last_trade_price c 61,816 38 11ms 46ms 266ms 4.6s 0 transport_latency_assumed last_trade_price e 61,199 61,198 78ms 5.2s 16.5s 33.0s 0 transport_latency_assumed last_trade_price h 67,121 7,055 30ms 614ms 5.1s 7.4s 0 transport_latency_assumed market_resolved b 1,494 13 51ms 363ms 1.2min 2.4min 0 transport_latency_assumed market_resolved c 1,542 160 5ms 3.6min 9.5min 9.8min 0 transport_latency_assumed market_resolved e 1,513 1,513 44ms 228ms 1.8s 1.0min 0 transport_latency_assumed market_resolved h 1,495 11 18ms 107ms 55.2s 2.4min 0 transport_latency_assumed new_market b 483 44 196ms 11.7s 48.7s 6.0min 0 transport_latency_assumed new_market c 483 52 33ms 2.9min 5.3min 10.7min 0 transport_latency_assumed new_market e 483 483 111ms 12.4s 59.4s 3.8min 0 transport_latency_assumed new_market h 483 103 82ms 12.5s 56.4s 6.9min 0 transport_latency_assumed price_change b 31,270,490 1,538,849 51ms 282ms 11.7s 31.5s 0 transport_latency_assumed price_change c 43,764,583 9,219,979 5ms 14ms 99ms 6.5s 0 transport_latency_assumed price_change e 36,583,716 36,583,436 44ms 247ms 11.1s 33.1s 0 transport_latency_assumed price_change h 44,657,756 15,279,030 18ms 36ms 831ms 7.5s 0 transport_latency_assumed tick_size_change b 2,968 not meas. 51ms 74ms 209ms 510ms 0 transport_latency_assumed tick_size_change c 2,808 not meas. 5ms 12ms 33ms 229ms 0 transport_latency_assumed tick_size_change e 3,016 3,016 44ms 108ms 413ms 6.9s 0 transport_latency_assumed tick_size_change h 2,968 not meas. 18ms 26ms 106ms 252ms 0 transport_latency_assumed
Captured and supplied are different numbers
Captured is how many rows each machine's own feed carried. Redundant capture means these OVERLAP, heavily and on purpose, so they do not partition the file and their sum is larger than it. Two machines each capturing nearly every event is the archive working, not double counting.
Supplied is how many rows of the merged file actually came from that machine. This one does partition: it sums to the file's row count exactly. But it is decided by the merge's tie-break, so a machine can capture almost everything and supply almost nothing. Read alone, either number tells a false story about the fleet, which is why both are printed and neither is called "the witness share".
What the skew columns mean
Arrival skew is the receipt timestamp minus the venue timestamp, in milliseconds. For six of the seven products the sidecar interprets it as transport latency, and marks that interpretation assumed: it rests on the venue timestamp being emit time, which we cannot verify from outside.
book is not latency and must not be read as latency. Its
venue timestamp is the last time the book changed, so the sidecar interprets it as book
staleness and warns in its own note that large values mean a quiet book rather than slow
transport. Its maximum runs to weeks for that reason. This is why the table above gives a
row per product per machine and never one figure for the hour: percentiles do not average,
and these seven do not measure the same quantity.
Check it yourself
The digest above is a claim. This is how you refuse to take our word for it:
curl -O https://archive.pendulumflow.com/v3/2026-09-03/21/2026-09-03T21.parquet sha256sum 2026-09-03T21.parquet # compare with the sha256 line in the report above
A range digest certifies bytes. That those bytes are the product the manifest names comes from the manifest, which is why both are published and both are checkable.