thanks andrew!

Out of curiosity, do you guys try to backfill when data is missing as well? I just saw that 9/27 is missing quite a few files: 0,7,10,17,21,22,23.

On Mon, Sep 21, 2015 at 6:25 AM, Andrew Otto <aotto@wikimedia.org> wrote:
Hi Tony,

Indeed!  We noticed this on Friday too.  A couple of changes[1] were recently made to improve a few things with our webrequest processing, and it seems things have greatly slowed since then.  We’ll talk about this problem for the first time today.  I’m not sure when things will get better, but I believe they eventually will.

[1] https://github.com/wikimedia/analytics-refinery-source/commit/fd59a13c4a6efbad333f83d7f0c47759faf8c488

Sorry for the delay!
-Andrew


On Sep 18, 2015, at 04:23, Tony Ho <tonyho1992@gmail.com> wrote:

Hi WikiMedia Analytics,

I'm a student who has been doing work with the page count files from wikimedia.

During the last few days, it looks like the latest page count is being published slower than before.

Usually, when I go to the following link:


I could see what happened an hour ago sometime within an hour or so after that.

Is this property still going to be true? This seems to not be the case for 9/16 and 9/17.

Also, pagecounts-20150916-090000.gz does not seem to be of correct size.

Thanks,

Tony Ho
_______________________________________________
Analytics mailing list
Analytics@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/analytics