Quarry seems much more intuitive and easier to use and share queries than Superset.

I am interesting in make tools with those data, but I am not familiar with the analytics infrastructure, I am more familiar with the Toolforge infrastructure, so in my idea those data could be in a MariaDB database in the tools.db.svc.wikimedia.cloud server, and the data could work like the wikis replicas, the complete data is restricted and the "safe" data is available in a public database (that ends with _p) through a view that hides the sensible data. Or, to be simpler, create only a public database with just the safe data.

Every time wmopbot restarts it restarts the k8s container. Maybe other bots reconnect without restart the process and container and that is preventing the changes to be applied to them.

wmopbot doesn't disconnect for more than 24h, the problem seems to has been resolved, thank you to all involved in the solution.

I asked a Libera staffer (moonmoon) and they said the traffic between Libera servers and Toolforge had a spike on october 10th that was not seen in other connections, after that the traffic didn't have a significant increase.

I also found that started on 2022-10-10, approximately at 13:50 UTC

Why I didn't receive any e-mail notification about that?

The cloak queue now does those checks and mark the invalid characters.

The filter that disallowed the edit does not disallow newcomers add images, the filter 161 disallow edits with some sign of inappropriate summary, often related with vandalism. One of the things that the filter disallow is summary with more than 20 consecutive characters without spaces, the summary that the image recommendation added is "/* growthexperiments-addimage-summary-summary: 1 */", that is why the filter disallowed the edit.

Fixed. I changed the code to not use JQuery.

The 15 missing pages have just appeared in the category list, I checked an hour ago and they were not in the list. It took 8 days after the edit to all pages be put in the list.

I checked the category list now, 5 days after the template was edited, and the majority of the pages are in the category list, I checked yesterday and they were not. I found 15 pages that still don't appear in the category, this is one of them.

I didn't find the total per month in those files, it will not be provided anymore? I have some tools that use the total pegecounts per month per article, that is the only data I need from the pagecounts files.

I made some queries to get some initial results:

I will make my script use the bz2 file when the uncompressed file is not complete.

For who is very new in Kubernetes, it is weird that you have to tell the system what programming language you are using, that not happen in grid engine. That is why I used the jsub command as reference, you can just run jsub foo.py with no arguments and it works, I probably not explained that well in the title, I am changing it. I didn't understand your comment about I use python in examples, python is the language I use more, but the command idea is to work to all languages that has a image compatible in kubernetes.

The idea is not create a script that always works for every application, but something that works for some simple applications, to help people start use kubernetes, or when we just want to make some test. The script can print the configuration and ask if you want to run with that config in order to avoid some errors. It can also have an argument to indicate that we just want the script to generate a yaml file, which we will reviwer, increment and run with kubectl, it is easier than create a yaml from scratch.

Thank you for find the bug. I made a change in the regex R2075:e8d36adee730 and that fixed the problem.

I know how to use style classes to do that, and I know many developers have restrictions about inline styles, but I find sometimes inline styles are better than classes and the editors can decide which is the best in each case. This task is not the appropriate place to talk about styles, so I will not extend on that.

The solution of this task would also help to create different content/styles for mobile version using, for example, {{#ifeq:{{SERVERNAME}}|en.m.wikipedia.org|you are using the mobile version|you are using the desktop version}}

Fixed

My concern about a general IRC name was that it could be used in tasks not related to IRC operators, like IRC meetings, IRC bots bugs or issues in irc.wikimedia.org. I don't know if we want to track those kinds of tasks.

IRC operators are like "sysops" and group contacts are like "stewards" in IRC. We usually discuss the IRC issues in #wikimedia-ops and #wikimedia-ops-internal between all operators.

The project will be used to tag tasks related to Wikimedia IRC channels operators and group contacts like T164651.

@tizianopiccardi, AlexZ unbanned *!*@*.ch, can you confirm if you can join channels now?