Hi,
due to gadolinium and its webstatscollector process having had load issues (bug 70118 [1]. Ottomata fixed the root cause already), the webstatscollector files at
https://dumps.wikimedia.org/other/pagecounts-raw/2014/2014-08/
for the hours between 2014-08-24 14:00 and 2014-08-27 21:00 might exhibit a higher loss than usual.
The files up to 2014-08-24 14:00 are ok. The files from 2014-08-27 21:00 onwards are ok. But the files in between still need closer examination. We'll track progress on that on bug 70118 [1].
If you consume webstatscollector files directly, or indirectly (stats.grok.se, wikistats, ...) please be aware of data issues for that period.
Sorry for the inconveniences, Christian
[1] https://bugzilla.wikimedia.org/show_bug.cgi?id=70118
Hi,
sadly, it's a bad week for webstatscollector:
Bug 70136 [1] - No usable webstatscollector output files for 2014-08-28 17:00 -- 21:00 Bug 70140 [2] - Webstatscollector counting HTTPS from ulsfo twice
Best regards, Christian
P.S.: Only relaying them again here, as webstatscollector data is public and has so many users, and those bugs might really affect its use. :-/
If bugs keep coming at that rate, we'll find a different means to announce them [3] to keep the noise low on the list.
[1] https://bugzilla.wikimedia.org/show_bug.cgi?id=70136
[2] https://bugzilla.wikimedia.org/show_bug.cgi?id=70140
[3] Like asking interested people to watch
https://wikitech.wikimedia.org/wiki/Analytics/Webstatscollector#Events_and_k...