Summary
For at least one package (clawmetry), the download stats on ClickPy look stale: the "last day" figure shows 0 and the total has been frozen at ~210K for ~2 days, even though the package is still being downloaded. The canonical PyPI download source is current, which points to a lag/stall in ClickPy's own ingestion rather than an upstream problem.
Steps to reproduce
- Open https://clickpy.clickhouse.com/dashboard/clawmetry
- Look at the Downloads by period card.
Observed (as of 2026-05-27)
ClickPy vs pypistats.org (both derive from the same PyPI file_downloads / Linehaul → BigQuery dataset, so they should broadly agree):
| Window |
ClickPy |
pypistats.org |
| last day |
0 |
1,745 |
| last week |
11K |
14.8K |
| last month |
71K |
33.9K |
| total |
210K (unchanged ~2 days) |
— |
The key signal is last-day = 0 on ClickPy vs ~1,745 on pypistats, and the total not moving for ~2 days.
Why this looks like a ClickPy-side ingestion lag
- pypistats.org is up to date (~1.7K/day), so the upstream PyPI BigQuery dataset is still receiving data — this rules out an upstream Linehaul/BigQuery outage.
- "last day = 0" + a frozen total is the classic symptom of the most recent day partition(s) not being loaded into ClickHouse yet.
This may be related to the data-completeness gaps noted in #210.
Questions
- Is the daily ingestion job currently behind / stalled?
- What is the expected lag between a download happening and it appearing in ClickPy, so a "0 last day" can be distinguished from a real stall?
Thanks for maintaining ClickPy — happy to provide more detail.
Summary
For at least one package (
clawmetry), the download stats on ClickPy look stale: the "last day" figure shows 0 and the total has been frozen at ~210K for ~2 days, even though the package is still being downloaded. The canonical PyPI download source is current, which points to a lag/stall in ClickPy's own ingestion rather than an upstream problem.Steps to reproduce
Observed (as of 2026-05-27)
ClickPy vs pypistats.org (both derive from the same PyPI
file_downloads/ Linehaul → BigQuery dataset, so they should broadly agree):The key signal is last-day = 0 on ClickPy vs ~1,745 on pypistats, and the total not moving for ~2 days.
Why this looks like a ClickPy-side ingestion lag
This may be related to the data-completeness gaps noted in #210.
Questions
Thanks for maintaining ClickPy — happy to provide more detail.