Console
The latest run on this page finished at 18:40 UTC on Monday 19 October. A run still going is not on this page yet.
Latest day
2026-10-19 ran 3 runs and published 195 of 240 planned items. 45 items failed.
3 runs are worth a look
Worst thing right now
26 feeds are resting, so nothing they carry reaches the digest. Each is asked again after 5 runs. On Pipelines.
Site size
45.7 MB of the 1 GB limit - room for about 386,000 more articles.
The days control above needs JavaScript, and so do the sections below - they draw rows a browser fetches. Their shape is 14 days.
The model spent 29 h 21 m of this. Summaries →
At a glance
Articles published
4677
- 19 Oct 2026
- Articles published
- 195
in these 14 days
Visuals published
0
- 19 Oct 2026
- Visuals published
- 0
in these 14 days
Articles published against visuals published over the same days.
Run health
The published day beside the runs that tried it.
Planned against published articles, and each run's own recorded outcome.
Articles published against planned
Planned and failed are totals across the day's runs. Published counts each article once, and only if it reached readers. So the line can be lower than the run squares suggest.
- 6 Oct 2026: 323 published of the 663 planned, 287 failed, share published 49%
- 7 Oct 2026: 395 published of the 550 planned, 69 failed, share published 72%
- 8 Oct 2026: 398 published of the 570 planned, 64 failed, share published 70%
- 9 Oct 2026: 400 published of the 501 planned, 41 failed, share published 80%
- 10 Oct 2026: 358 published of the 488 planned, 32 failed, share published 73%
- 11 Oct 2026: 357 published of the 411 planned, 36 failed, share published 87%
- 12 Oct 2026: 235 published of the 328 planned, 22 failed, share published 72%
- 13 Oct 2026: 249 published of the 520 planned, 264 failed, share published 48%
- 14 Oct 2026: 393 published of the 543 planned, 56 failed, share published 72%
- 15 Oct 2026: 390 published of the 461 planned, 71 failed, share published 85%
- 16 Oct 2026: 325 published of the 400 planned, 75 failed, share published 81%
- 17 Oct 2026: 337 published of the 477 planned, 91 failed, share published 71%
- 18 Oct 2026: 322 published of the 395 planned, 73 failed, share published 82%
- 19 Oct 2026: 195 published of the 240 planned, 45 failed, share published 81%
- 19 Oct 2026, the newest day
- Planned
- 240
- Published
- 195
- Failed
- 45
- Share published
- 81%
- Run 1
- 67 of 80 succeeded, 13 failed
- Run 2
- 69 of 80 succeeded, 11 failed
- Run 3
- 59 of 80 succeeded, 21 failed
Point at a day to read its counts and every run on it. Left and Right step through the days, Escape returns to the newest.
Each square is one run, with run 1 at the bottom. A run is judged only on the articles it tried. Red: fewer than 70% of them succeeded, or the run failed. Amber: at least one of them failed, the run had nothing new to try, it reused yesterday's list of sources, or another run's failed article was still missing. Green: none of these. A skipped article, already published or repeated by a feed, never counts.
What one more article costs
2,031 B an article.
Each day's bytes per article against the window's median; give or take 26 B.
How long we can keep publishing under the 1 GB Pages cap. Bytes the committed payload tree gained on each published day, over the articles that day published. Over 14 days.
At 2,031 B an article and a median of 347 articles a published day, the cap has room for about 504,985 more articles: about 4.0 years at this window's rate. The cap measures the built site, which is larger than the payload tree, so this is the most room we have.
Up 8.4 MB over 14 days.
- 19 Oct 2026, the newest day
- Payload bytes per article
- 2000.7179487179487
Point at a day to read it. Left and Right step through them, Escape returns to the newest.
Nothing changed about how the summaries are written inside these 14 days.
- 2026-10-06: 2101 B per article, outside the band.
- 2026-10-07: 2009 B per article.
- 2026-10-08: 2055 B per article.
- 2026-10-09: 2034 B per article.
- 2026-10-10: 2036 B per article.
- 2026-10-11: 2047 B per article.
- 2026-10-12: 2043 B per article.
- 2026-10-13: 2063 B per article, outside the band.
- 2026-10-14: 2019 B per article.
- 2026-10-15: 2005 B per article, outside the band.
- 2026-10-16: 2009 B per article.
- 2026-10-17: 2029 B per article.
- 2026-10-18: 2022 B per article.
- 2026-10-19: 2001 B per article, outside the band.
What is failing, by stage
Each failed stage against the other stages in the same day's stack, not a rate against another day's volume.
The failure record is still loading.
The query has no article identity; this counts rows, not distinct articles.
Where an item's time went
Whether a run is getting slower on the same work
Time per item by stage
What one item cost the model
Where the run's time went, on the run's own clock
Visuals drawn for articles
0 visuals published.
The median day has no minutes on record over these 14 days, and puts a visual on 0% of what it published, below the 5% floor.
Visuals published against articles that reached drawing.
Over 14 days with the chart-only gate on, chart drawing is retired above 6 minutes per published visual or below 5% of the items carrying a published visual.
Minutes per visual -
No minutes are on record.
Published articles with a visual 0%
Retired below 5%, on the median day.
- 19 Oct 2026
- Articles with a visual
- 0%
Point at the line to read a day.
Every share is of the items the visual planner reached.
Show these figures day by day, over these 14 days
One row per day in the open window, newest first. A dash means no minutes are on record.
| Day | Reached | Asked the model | Visuals drafted | Visuals published | Items published | Minutes spent | Minutes per visual |
|---|---|---|---|---|---|---|---|
| 2026-10-19 | 195 | 195 | 57 | 0 | 195 | 0.0 | - |
| 2026-10-18 | 322 | 322 | 65 | 0 | 322 | 0.0 | - |
| 2026-10-17 | 386 | 386 | 100 | 0 | 337 | 0.0 | - |
| 2026-10-16 | 325 | 325 | 78 | 0 | 325 | 0.0 | - |
| 2026-10-15 | 390 | 390 | 75 | 0 | 390 | 0.0 | - |
| 2026-10-14 | 487 | 487 | 96 | 0 | 393 | 0.0 | - |
| 2026-10-13 | 251 | 251 | 56 | 0 | 249 | 0.0 | - |
| 2026-10-12 | 306 | 306 | 77 | 0 | 235 | 0.0 | - |
| 2026-10-11 | 375 | 375 | 85 | 0 | 357 | 0.0 | - |
| 2026-10-10 | 456 | 456 | 89 | 0 | 358 | 0.0 | - |
| 2026-10-09 | 460 | 460 | 107 | 0 | 400 | 0.0 | - |
| 2026-10-08 | 506 | 506 | 143 | 0 | 398 | 0.0 | - |
| 2026-10-07 | 481 | 481 | 122 | 0 | 395 | 0.0 | - |
| 2026-10-06 | 376 | 376 | 83 | 0 | 323 | 0.0 | - |
What the extractor found
Chartable articles against narrative and unclassified ones, beside recorded extraction flags.
Of those, published without one
100%
2487 of 2487 chartable published articles
The record does not identify which extraction checker ran or its version. A false flag is the record's default, not proof that an extraction check ran.
Whether the yield is falling
- 2026-10-19, the newest day
- Enough figures to draw
- 99
- Published carrying a chart
- 0
Point at a day to read it. Left and Right step through them, Escape returns to the newest.
Nothing changed about how the summaries are written inside these 14 days.
- 2026-10-19: 99 enough figures, 0 carrying a chart.
- 2026-10-18: 162 enough figures, 0 carrying a chart.
- 2026-10-17: 191 enough figures, 0 carrying a chart.
- 2026-10-16: 155 enough figures, 0 carrying a chart.
- 2026-10-15: 204 enough figures, 0 carrying a chart.
- 2026-10-14: 228 enough figures, 0 carrying a chart.
- 2026-10-13: 241 enough figures, 0 carrying a chart.
- 2026-10-12: 130 enough figures, 0 carrying a chart.
- 2026-10-11: 155 enough figures, 0 carrying a chart.
- 2026-10-10: 176 enough figures, 0 carrying a chart.
- 2026-10-09: 205 enough figures, 0 carrying a chart.
- 2026-10-08: 273 enough figures, 0 carrying a chart.
- 2026-10-07: 257 enough figures, 0 carrying a chart.
- 2026-10-06: 199 enough figures, 0 carrying a chart.