TOLL

What software pays for, recorded every six hours.

Snapshot
2026-09-15 16:57 UTC
Sweep
daily-2026-09-15 · finished 2026-09-15 07:10 UTC
Listed
16,167

Status

What the collector did and when. Every registry pull, every sweep, every restart and every alert is a row, so a gap in the record always has a stated cause.

this page, as of 2026-09-15 17:40 UTC · method · liveregenerates every five minutes
Last successful registry pull
2026-09-15 16:57 UTC
as of 2026-09-15 16:57 UTC · method · live · pulls run at 17 past every sixth hour UTC; next due 2026-09-15 18:17 UTC
Last finished daily sweep
daily-2026-09-15
as of 2026-09-15 07:10 UTC · method · live · planned 15,523, completed 15,708, 203 minutes; the daily sweep is due at 03:47 UTC
Completed exceeds planned by 185: planned is counted once when the first chunk starts, and endpoints listed by a pull during the day join the same sweep, so the difference is the day's new listings that were swept.
Sweep runs left open
10
as of 2026-09-11 09:24 UTC · method · live · stopped part-way through a chunk and not writing since; 5 others are open by design, each continued by the next run
A new-listings run is closed only when its last chunk clears the queue, so a run that filled its chunk stays open on purpose. These did not: the worker stopped part-way and restarted, which opens a fresh run rather than resuming the old one, and the abandoned row stays. The endpoints they had not reached were swept by the run that followed, so the record has no hole; the rows do.
Quality snapshots taken
28
as of 2026-09-15 16:57 UTC · method · live · 412,444 rows, never deleted
Registry pulls, all time
56 ok · 2 failed
as of 2026-09-15 16:57 UTC · method · live · Binance since 2026-09-10 19:08 UTC; Coinbase since 2026-09-10 19:08 UTC

Registry pulls, most recent first

One row per registry per run. A pull outside the six-hour schedule is a manual run.passfail

Recent registry pulls with their timing, record counts and outcome
started, UTCregistrysecondsrecords pulledregistry reportedrequestsoutcome
2026-09-15 16:58 UTCBinance598098010pass
2026-09-15 16:57 UTCCoinbase1916,04316,04317pass
2026-09-15 11:43 UTCBinance498098010pass
2026-09-15 11:42 UTCCoinbase2215,85315,85316pass
2026-09-15 04:53 UTCBinance598098010pass
2026-09-15 04:52 UTCCoinbase2315,57715,57716pass
2026-09-14 21:49 UTCBinance498098010pass
2026-09-14 21:48 UTCCoinbase2315,79815,40017pass
2026-09-14 12:44 UTCBinance498098010pass
2026-09-14 12:43 UTCCoinbase2515,38215,38216pass
2026-09-14 04:54 UTCBinance398098010pass
2026-09-14 04:54 UTCCoinbase2315,38515,38516pass
2026-09-13 20:49 UTCBinance498098010pass
2026-09-13 20:48 UTCCoinbase1915,35215,35216pass
2026-09-13 16:15 UTCBinance498098010pass
2026-09-13 16:15 UTCCoinbase2215,18915,18916pass

Sweeps, most recent first

Daily sweeps cover every listed endpoint; “new” sweeps probe listings first seen at a pull, in chunks of 200, each chunk its own run. Planned is counted when a run first starts, and endpoints listed later in the day join the daily sweep, so completed can exceed planned. A run that is due shows as due; a run that is late shows as a failure of the collector, in the same table.passunobservedmalformedfail

Recent sweeps with their plan, completion and outcome
runkindstarted, UTCfinished, UTCplannedcompletedstopped becauseoutcome
new-2026-09-15T17-01-37-864Znew2026-09-15 17:01 UTC2026-09-15 17:02 UTC8181pass
new-2026-09-15T16-58-41-653Znew2026-09-15 16:58 UTCstopped at the 200 chunk, continued by the next run281200pass
new-2026-09-15T11-46-48-266Znew2026-09-15 11:46 UTC2026-09-15 11:48 UTC8585pass
new-2026-09-15T11-44-05-153Znew2026-09-15 11:44 UTCstopped at the 200 chunk, continued by the next run285200pass
new-2026-09-15T04-54-30-491Znew2026-09-15 04:54 UTC2026-09-15 04:56 UTC191191pass
daily-2026-09-15daily2026-09-15 03:47 UTC2026-09-15 07:10 UTC15,52315,708pass
new-2026-09-14T21-49-38-378Znew2026-09-14 21:49 UTC2026-09-14 21:52 UTC196196pass
new-2026-09-14T12-45-19-314Znew2026-09-14 12:45 UTC2026-09-14 12:45 UTC4141pass
new-2026-09-14T04-56-17-031Znew2026-09-14 04:56 UTC2026-09-14 04:57 UTC7777pass
daily-2026-09-14daily2026-09-14 03:47 UTC2026-09-14 06:36 UTC15,47615,522pass

Snapshot cadence

Hours between consecutive quality snapshots. The schedule is six; a shorter gap is a manual pull, a longer one a delayed or skipped scheduled run.

Recent quality snapshots and the hours since the previous one
snapshot, UTCendpoints with countershours since previous
2026-09-15 16:57 UTC16,0395.3
2026-09-15 11:42 UTC15,8496.8
2026-09-15 04:52 UTC15,5737.1
2026-09-14 21:48 UTC15,3959.1
2026-09-14 12:43 UTC15,3787.8
2026-09-14 04:54 UTC15,3818.1
2026-09-13 20:48 UTC15,3484.6
2026-09-13 16:15 UTC15,1844.4
2026-09-13 11:49 UTC14,5717.1
2026-09-13 04:45 UTC14,6478.0
2026-09-12 20:44 UTC14,6090.1
2026-09-12 20:40 UTC14,6090.1

Alerts

Every alert the collector raised, collapsed by message, and every crash-loop pause of the sweep worker. What happened is published; the driver text behind it is not.

Alerts and crash-loop pauses, most recent first
last at, UTCwhattimesfirst at, UTC
2026-09-14 19:29 UTCalert: worker exited12026-09-14 19:29 UTC
2026-09-14 01:57 UTCalert: worker exited12026-09-14 01:57 UTC
2026-09-13 08:43 UTCalert: worker exited22026-09-11 19:40 UTC

Worker starts

36 starts of the sweep worker on record, the first at , the latest at . A start is normal after a deploy or a laptop wake; the crash-loop guard counts starts in a ten-minute window and pauses the worker when there are too many, and a pause is listed above.

Every recorded start, 36
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 4 start(s) in the last 10 min
  • · 4 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min
  • · 1 start(s) in the last 10 min
  • · 4 start(s) in the last 10 min
  • · 2 start(s) in the last 10 min

Performance baseline

Lighthouse on an endpoint page already in the cache, mobile emulation, slow 4G, four-times CPU slowdown. Every column names its method, because a reading is not comparable with one taken differently. The 13 September reading was taken before the redesign and its own method was not recorded, so it is not a like-for-like pair with any column beside it. Lighthouse 12.8 was then run twice on the strip-chart recorder and twice again on the posted board that replaced it, once under simulated throttling, its default, and once under applied. The budget is held to the applied readings, which are the measured ones; under simulation the largest contentful paint fails the same budget in both worlds, and all four stand here. The board adds three server-rendered figures and a second font, which is the 29 KiB between the two transfer figures.

Lighthouse baseline figures and their budgets
measure13 Sep, method not recorded14 Sep, recorder, simulated14 Sep, recorder, applied14 Sep, board, simulated14 Sep, board, appliedbudget
Performance score9297989899
First contentful paint1.2 s1.0 s1.7 s0.9 s1.6 s
Largest contentful paint1.5 s2.6 s, over1.7 s, under2.4 s, over1.6 s, underunder 1.8 s
Cumulative layout shift00, under0, under0, under0, underunder 0.05
Total blocking time300 ms80 ms120 ms40 ms70 ms
Transferred237 KiB256 KiB256 KiB285 KiB286 KiB

The site ships no client script beyond the framework's deferred bundle and one small control that fills the access date in citations. The budgets and the framework decision are recorded on the method page.