I support @MBinder_WMF's request.

https://gerrit.wikimedia.org/r/460410 hasn't been merged and is V:-1.

In T203814#4576030, @Nuria wrote:

After talking with @Ottomata on irc about the reasons why eventlogging is not well suited to do error logging I have written a wikitech page on this regard: https://wikitech.wikimedia.org/wiki/Analytics/Systems/EventLogging/NotErrorLogging#Eventlogging_is_not_Well_suited_to_do_error_logging

Per T177219#4493965, this is no longer needed.

In T199116#4597455, @zeljkofilipin wrote:

As far as I can see it, this should be done in Quibble. I'm not familiar with it, so it's unlikely I would be working on this.

All subtasks have been resolved. Being bold.

@alexhollender: A friendly heads up that, per T203992#4579032, this change will be live on all wikis next Thursday (2018/09/20) and so should be reviewed by you. You can see the change in action here: https://en.wikipedia.beta.wmflabs.org/wiki/Special:Preferences#mw-prefsection-rendering

In T204143#4579995, @Jdrewniak wrote:

On http://readers-web-master.wmflabs.org however, ReadingDepth doesn't seem to be logging at all, leading me to believe that maybe the WikimediaEvents repo isn't installed there. @Niedzielski can you confirm that? (I wasn't able to log into the server myself :/ )

Will this change make it even easier to extract the time module to a library that can be used by other teams.

Following on from @ovasileva's comment above, we (Web) also agreed that there are enough known unknowns (where the code lives and how we post-analysis requests) to warrant an estimate of 13 points.

All subtasks of this task are resolved. Being bold.

@pmiazga: Could you provide instructions on how to test this?

🤘

@Bawolff: How's this going? Sorry if I've missed an update elsewhere.

OTOH the Google Rich Cards extension seems to add these script tags already (without the sameAs property AFAICT).

^ Formalising that this blocks T202748 and T203013.

In T198946#4556902, @Jdlrobson wrote:

To me, rather than creating a new extension this feels like something that should live in one of the Wikibase extensions. Can we reach out to Wikimedia DE to see where would be appropriate?

Quoting from @dr0ptp4kt's notes of the meeting on Wednesday, 29th August:

I can confirm that I see the same behaviour described in T118063#4547350 – it's always good when browsers behave as specified!

In T203013#4543477, @Krinkle wrote:

Also, if all we need is to differentiate a handful of events beaconed from the same page view, we may want to consider a JavaScript number that increments. Would that suffice?

Note: be careful as page previews also has a concept of a linkInteractionToken for which there is no shared/well-defined API.

☝️ per @mobrovac's review of https://gerrit.wikimedia.org/r/#/c/mediawiki/services/chromium-render/+/451817/.

I would go so far as to say that this is UBN!

In T137449#4536073, @Jdlrobson wrote:

A few suggestions:

• TextExtracts, Monobook and Vector should be made security-only. Vector we might want to clarify we review tasks related to current goals.
• WikidataPageBanner, QuickSurveys - currently 7 days but we've not been doing that. Maybe bump to 14.

👍

In T202026#4528900, @Jdlrobson wrote:

I've defined https://meta.wikimedia.org/wiki/Schema:ClientSideError and I've made a few editorial changes to the names provided in the description.

I met with @Jdlrobson about this after today's kickoff ritual. We tested this in production (enwiki), and proved to ourselves that the fix works by following the QA steps in the description with the window set to 450px and 375px widths.

Thanks for following up, @Jdlrobson.

In T199000#4523410, @Jdlrobson wrote:

When JS is disabled we don't show the icons. This was the case before the patches and now. We can revisit this now as icons can be shipped as uris rather than data uris but out of scope for this patch.

I tested this on the BC briefly and saw:

@ovasileva: Since you're the PO and you've just QA'd this, who do you think should sign it off?

Hrrm. I'm seeing the same thing that @nray saw in T200518#4467436 on both reading-web-staging and readers-web-master, i.e. not what @alexhollender saw in T200518#4499137.

I'm going to leave this in Needs QA for now until we figure out if the horizontal padding of the wikitext editor textarea has indeed changed.

Following on from my comment above, I've noticed a couple of differences in the treatments of the last modified bar with JavaScript enabled and disabled.

Search icon in mobile view and search icon inside Special:Search/SearchOverlay

Claiming this per Readers Web's discussion about clearing down the Needs QA column during yesterday's standup.

🎉🎉🎉 Excellent work, y'all!

Create a new ErrorLog schema.

@mobrovac/@Pchelolo: Could you review rMSCR3e074fd4abf8: Handle the stopping chromium rendering correctly and rMSCR130e32ee19dd: To prevent resource starvation, try to kill chromium after 3s at your earliest convenience?

Quoting @Tgr from IRC:

In T201954#4502632, @Jdlrobson wrote:

Prioritise as necessary!

Notes from yesterday's standup ritual:

Further to the above, AIUI both T181623: Chromium-render doesn't handle browser connection abort well and T177765: Security review of mediawiki-services-chromium-render block Proton's deployment. @pmiazga has submitted two changes for the former and @charlotteportero has just let me know that the latter is being worked on and is expected to be resolved within two weeks.

Regarding T181623#4495059, can anyone in Readers Web review @pmiazga's changes? I'll also ask during today's standup ritual.

@Tbayer: Did you want to set up a meeting with an engineer to go through the instrumentation manually (further to the QA steps that @Jdlrobson added). IIRC we did that for Page Previews and it was useful.

Sorry if we have and I've missed the comment but have we determined that the spike that @Jdlrobson previously observed (see below) has flattened out?

In T201068#4476098, @mobrovac wrote:

Could you elaborate on that? There would be validation upon event message intake (i.e. that they conform to a particular schema).

Being bold… again!

🎉🎉🎉

I stand corrected. Especially about the note about the control group, which I think I confused in the construction of the Page Previews A/B tests when we were running them. Thanks, @Tbayer!

I didn't mean to cause confusion with my comment. If I have, then I apologise.

As an engineer, I want good client libraries to produce events so that I don't have to write them myself.

In T200792#4475276, @Jdlrobson wrote:

An A/B test by definition will split users into 2 equal buckets - 50% group A and 50% group B. The sampling rate defines the total number of users in the test,soif we are using a sampling rate of 10% of 100 users, that will be 10 users, with 5 in group A and 5 in group B.