How will the SDK get the values of the contextual attributes in the Android environment? The PHP pattern may not necessarily be the idiomatic implementation for Android. Have you talked this part through with @Dbrant ?

Until this gets used (an experiment and instrument are run in the Android app) this should remain a v0.1 -- let's consider that when we think through documentation

Thanks for filing! We are interested in striking a balance between ensuring that developers can debug their instrumentation code as needed, and creating too much noise. Do you have any particular suggestions around how to balance this tension?

Data will be unified and put into a single iceberg table as part of SDS2.2.6

@Kgraessle should we mark this as closed? The experiment completed, and we decided to not move forward with the change we tested.

T409913: Create PHP equivalent of mw.testKitchen.getInstrument() is doing the SDK side of this work, which will simplify this task to be UI specific.

Architecture, including Flows: https://wikitech.wikimedia.org/w/index.php?title=Test_Kitchen%2FArchitecture

@cjming I noticed on the Architecture page that there is a s[[ https://wikitech.wikimedia.org/w/index.php?title=Test_Kitchen%2FArchitecture&diff=2271265&oldid=2260795 | ection for the GrowthBook JS SDK ]]-- any reason why? Should this be maintained? It's not really part of our system, is it?

(And changed Glossary into sections, as it was getting hard to follow alphabetically) https://wikitech.wikimedia.org/w/index.php?title=Test_Kitchen%2FGlossary&diff=2373110&oldid=2373106

Glossary Section: https://wikitech.wikimedia.org/w/index.php?diff=2373106

FYI - agent_ua_string is now available as a contextual attribute https://phabricator.wikimedia.org/T385180

@ssingh - when might your team be comfortable with us running the test? Has the incident been resolved or mitigated?

Happy to help @AStein-WMF if you have any questions about Test Kitchen. Product teams are using it to measure impressions, page views, etc so it could be a good like-like comparison with e.g. trends in readership.

If I pass the mpo parameter but am not enrolled, do I generate events?

Marking this as low for right now - as we approach SDS2.3 Growthbook for Configuration, I expect we'll learn more about whether the Test Kitchen UI persists, or if GrowthBook makes it easier/harder

Is creating an experiment enough? Do we need a step to "turn on" at all? Could we default to "ready" when saved? I reached out to platform users with:

This came up again with @DLynch and editing -- boosting to high

@phuedx and this would be particularly bad for Growth's future experiments (Revise tone structured task, slated for Jan), correct?

I have now archived the Metrics Platform board

I just tested this - video shows that:

• The sample rate defined for "All wikis", the value does not propagate to any subsequently added rows
• The sample rate fields in newly added rows for specific wikis are empty, and require users to input a sample rate manually

Thanks for the context @HFan-WMF ! Totally understandable that it happened, and I think that stepping back to ask "how might we have caught this" is always a good step in trying to prevent it in the future. Lots of unique elements on this one, so no real golden solution.

Oh no! Any idea why we deployed on mobile without instrumentation? Or how we missed the page-visited event name? I'm thinking more about refining our process so as not to miss these in the future.