Approved! (airflow-analytics-product-admins membership)

Approved! (airflow-analytics-product-admins membership)

@AndrewTavis_WMDE asked me for some thoughts/suggestions here :)

a simple way to accomplish this

I think there's a simple way to accomplish this but I don't think the end result would be particularly useful. I believe that for the end result to be useful "I want to know how many users have this feature enabled" this will need careful planning and consideration to account for the complexity of user preferences.

@VirginiaPoundstone: Howdy! The underlying dataset will probably be the hardest part of this because of the challenges of how user preferences are stored and used. And then yeah, a Superset dashboard would be the simplest way to make that data available to the end users. It wouldn't be through Turnilo because the metrics aren't additive across dimensions, so it would need to be Superset.

In T252227#9655162, @dr0ptp4kt wrote:

Okay, if I understand correctly, then the idea would be to...

1. Continue "allowing" tagging of wprov for non-200 HTTP responses. It's mainly important people don't accidentally count those as pageviews when they're not pageviews (i.e., they should be using is_pageview or something similarly precise). It's useful to be able to quickly zoom in on these sorts of requests anyway, so even for a 30x response it is nice to have.

2. If there's a 30x response for a redirect from desktop to mobile web and the URL came bearing a wprov, add that same wprov parameter name-value pair and also add the parameter name-value pair of rprov=1 in the target redirect URL (that's the thing that will be emitted in the Location: header).

Do I understand correctly?

@VirginiaPoundstone Howdy! This is not a blocking anything. Thanks for checking!

Prompted by confusion in Slack

@Pginer-WMF: Please keep in mind that this curiosity needs to be prioritized in the context of other work requested of KC for Language. For him, the highest priority tasks are:

• T356765: Correlation between article length, number of translations within a time period, experience of users, and deletion rate.
• T341185: Design instrumentation approach for machine translation of Wikipedia article contents

So he might not get around to this for a while.

@KSarabia-WMF: You will probably need to submit a request through L3SC for this.

@nettrom_WMF will be developing it as an essential metric in the next FY for SDS 2.2 and will likely use that definition.

Assigning to Megan who will work on this in Q4 to get baselines & reasonable target in before start of next FY.

From https://www.mediawiki.org/wiki/Growth/Personalized_first_day/Newcomer_tasks/Experiment_analysis,_November_2020#Detailed_findings

Thank you so much for looking into it, @phuedx!!!

@jwang: Hive has been / is being deprecated. I recommend you use Spark SQL instead (which also works with Iceberg tables, which hive CLI and wmfdata.hive do not).

After examining:

• mediawiki.web_ui_actions stream config
• Schema for the MP-based instrument: /analytics/mediawiki/product_metrics/web_ui_actions/1.0.1
• The (desktop) instrument which makes a few modifications (like storing name as action_source) before submitting the interaction data with Metrics Platform.

I think it depends on the technical specifics of the migration, which @KSarabia-WMF would be able to verify. Essentially my main concern was potential presence of new permalink-copied events coming from original instrument that might not be present in the new MP-based instrument, which would impact the data QA. If the new instrument automatically produces these new permalink events, then all is good and there's no concern.

Hey folks, is this being coordinated with the Web team who are currently in the middle of migrating *UIActionTracking to the Metrics Platform? T344274: Adopt Web Team Instrumentation to Metrics Platform

Queries done: https://gitlab.wikimedia.org/repos/product-analytics/data-pipelines/-/tree/main/citation_needed/searches

Assigning to @Iflorez to perform the last step (as owner of the dashboard)

+1 to Isaac's proposed solution of carrying wprov forward as wprov but also setting rprov=1 in case of a redirect to simplify analysis.

Thank you @cmooney for taking this non-standard case on and helping KC out! This dual account thing has become a real thorn for KC so I'm glad we're on a path to get it taken care of.

Approved from my side, both for the request in general and analytics-product-users membership :)

I would advise against using "session" in the name. When Jason and I were writing https://docs.google.com/document/d/100B4c1GqHHCAGnWLbDrgMQIJzxNI8vuf7jEMKG7DKeg/edit?usp=sharing we surveyed the landscape of schemas/instruments and saw that everything was just a session and there being multiple levels of sessions: https://docs.google.com/document/d/11xTwL_j0BWgfdtZ_GOIlxg22rLP2lRraaA2c_-uMwJQ/edit#

Okay, so it's been a few years now and this bug still exists and impacts the quality of our analyses substantially (especially for Future Audiences experiments that are aimed at mobile users) and we're not any closer to instrumenting pageviews.

As with the other task, Irene has a to-do this week to wrap this up by updating https://meta.wikimedia.org/wiki/Research_and_Decision_Science/Data_glossary#Contributor_metrics

Irene has a to-do this week to wrap this up by updating https://meta.wikimedia.org/wiki/Research_and_Decision_Science/Data_glossary#Contributor_metrics

@SBisson: root-level dt is the client-side timestamp, while meta.dt is the server-side timestamp. Both are useful to have for analysis.

On contribution side:
In some cases the instrument sends event data containing activity_session_id and dt fields which are not present in version 1.2.0 of the contribution schema: https://gerrit.wikimedia.org/r/plugins/gitiles/schemas/event/secondary/+/refs/heads/master/jsonschema/analytics/mediawiki/wikistories_contribution_event/1.2.0.yaml so a lot of events aren't passing schema validation.

Oh that's brilliant! Thanks so much for looking into it and shining light on this @phuedx!

The decision has been made (to go with Vue/Codex).

Thank you, Connie!

Unsure to what end this request was for.

The ChatGPT plug-in experiment has concluded. We'll just need to do some clean-up (Phab task coming later).

https://docs.google.com/document/d/1rzD_rJEzH3HmyIekmtojfiwyXqPVdEIHBviFA8_HgH0/edit#heading=h.w7kdor9r67qo is as final as it will get. We'll need to adopt it widely on the team (and do more peer reviews in the first place) but right now there's nothing left to do here.

@Mayakp.wiki: you and others are already in product-analytics-announce@, but the alerts aren't sent to that.

Just tried the query in the description with some recent dates but the dates I picked didn't have any requests with '-' UA strings, so it's hard to know if the problem persists. The query runs fine with user_agent = '-' just returns no data. I just made a request to https://www.wikidata.org/wiki/EntitySchema:E1 (according to https://www.wikidata.org/wiki/Help:Namespaces namespace 640 is EntitySchema so I just picked one):

@Jelto: This is to access dashboards in Superset that do access private data, no need for Hadoop/analytics client servers access. So no SSH public key needed. See https://wikitech.wikimedia.org/wiki/Analytics/Data_access

@cchen: How about Superset & Hue?

Tagging DPE SRE in case this is specific to those tools.

Not yet. I believe @AndrewTavis_WMDE will be sharing some findings from WMDE side soon.

Thanks for merging & deploying, @JAllemandou!

Thanks so much @MoritzMuehlenhoff!!

@BTullis: heads-up that @cchen won't be able to until T356645: Production data & systems access restoration for Connie Chen is resolved

@Pginer-WMF: Does this question relate to any of the hypothesis work your team is doing? If so, can you please share with hypothesis?

@VirginiaPoundstone @WDoranWMF I renamed this task to refer to the (now deprecated) monoschema version of Metrics Platform. The partial migrations of the instruments were removed (see T351337, T351335).

Closing this as the partial migrations have been decommissioned (T351335, T351337) following the learnings from T340702 and re-architecture of the Metrics Platform.

Shay met with me for consultation on this.

Before closing this task / withdrawing the request I'd like to get a confirmation from Goran whether the level of access is still needed and if it's just for the WMDE pipelines, in which case it would make sense to prioritize migrating those to WMDE's Airflow instance so that we can eventually revoke the unrestricted access to highly sensitive information.

I'm also sorry for making the request without knowing about the prior request.

We should add expiry_date and expiry_contact fields to reflect the NDA

Thank you @AndrewTavis_WMDE for alerting us of this. Pinging @Manuel for visibility.

This leads me to conclude that use of the MediaWiki core JS HTML docs is minimal, while interest in the docs may be high.

@nettrom_WMF Thank you for sharing that code! I recently used it in T353666 and it was very helpful! Just wanted to show my appreciation.

@KCVelaga_WMF Sam and I just chatted about this and enwiki might actually be good enough as a proxy for the service in general, and that's unlikely for the service to break on one particular wiki and not others.