@Isaac The report and the content changes are live. Note: It may take a few hours for the cache to refresh.

1. Debug URL

@leila Should these ship with the next report, or wait and go out with the site upgrade?

@soworu https://www.mediawiki.org/wiki/Product_Analytics/Superset_Access

@soworu >7K responses. Response rate ~63%. I'm undeploying.

@soworu I'm not monitoring the survey responses, but I may still have access to them via Superset, so I'll check. What number would you consider a sizeable amount?

Survey is live. @soworu Should I take it down at June 23 or wait until it collect enough responses?

@soworu Gotcha! Will deploy at 20:00 UTC.

@soworu Q06 and Q09?

Randomization (Q06, Q08, Q09): yes, let's keep them randomized. The one thing to watch is if any of those has a catch-all option ("None of these" / "Other" / "Don't know") that really ought to sit at the bottom. If so, flag it and we'll set just that question to fixed order instead.

Please review the survey at https://en.wikipedia.org/wiki/Wikimedia_Foundation?quicksurvey=mobile-app-awareness-2026-en-survey&uselang=en.

1. Debug URL

@soworu A few limitations of the QuickSurveys extension:

@YLiou_WMF @soworu Yes. The doc is finished, right? I will do a pre-deploy.

@MJahnavi-WMF The report says Jan 2026, is that the right date?

Why use the Enterprise API over the REST API

I've removed ~50 dead links from the site. They are documented on the attached file.

@leila Here is the preview for the new Research landing page: https://wmf-research-preview.pages.dev/

Thanks @Jelto!

@Volker_E @Dzahn heads up on the current state after the upgrade.

@Jelto I've deployed design-landing-page and cleaned refs on deployement-charts for design-strategy. The new site is showing on https://design.wikimedia.org but at https://design.wikimedia.org/strategy it still shows the old site. I think design-strategy must be destroyed/uninstalled manually on helmfile/helm but not sure or able to do so.

In the past I calculated the response ratio (clicks / views) of some surveys using Superset. It varied wildly from survey to survey and language to language but typically it was ~0.5% I think. We can't properly estimate the responses but the unique visitors stats from PT: 1.87M/day; TR: 1.1M is good enough for the purpose of setting an initial value for the survey coverage as it can be adjusted later and/or the survey left to run for a few days less or more. That gives 3.1% for PT and 5.2% for TR but 7% and 10% is also a good start considering how noisy the estimation is. Currently on the Airtable Dashboard we have 851 Portuguese speakers and 77 Turkish speakers.

@leila I expect to have a near-final preview within three weeks and will proceed with items (2) and (3). I’ve created a new task to track the upgrade.

In T415725#11563856, @Wargo wrote:

In T415725#11563488, @brennen wrote:

I've seen it claimed that ?action=purge for a given page fixes the error, but @Lucas_Werkmeister_WMDE says not always, so maybe people are seeing a coincidence with something dropping out of cache?

Yes, it (or null-edit) helps only once for a page and after revisit (same user or another) the error returns.

In the context of T311945, I’m migrating the Design site to a Codex + VitePress stack. Next, I plan to migrate the Research site, reusing the same theme, since the two sites have similar content structures.

Since the site repositories have been migrated to GitLab and a new deployment infrastructure has been put in place, this task is no longer relevant.

@leila https://research.wikimedia.org/report.html?date=2025-12-18

@leila I've made the adjustments.

All ported. Preview:

Apologies for that. At the time of deployment the canary check was fine but I can see the spike on Logstash now. I believe I should've observed it further.

1210729 pre-deployed the non-enwiki surveys with 0 coverage.

Debug URL

@TAndic I can try to deploy it on Monday afternoon backport window and otherwise on the late window. Anyway it wouldn't be useful for this survey as I'm deploying it on Monday as well.

In T410696#11396160, @TAndic wrote:

@YLiou_WMF I checked the skip patterns for the survey and all looks fine on this end.
Methodological note: temp users (previously IP editors/anon) are now categorized with user/editors (see QuickSurvey documentation). If you see a drop in previous editing experience, this may be a result of the population change. (Alternatively, we could target all and then use the userInGroup setting to only display to temp for users, but having never used this and working on a short timeline, it may be risky as I believe there are still some open questions about the temp group functionality.)

@YLiou_WMF I'm available for deployment on Monday.