@Jdlrobson

are not captured in the same schema

what exactly does this mean? will it affect our assumption of "more people click on language button than hamberger menu"?

@phuedx: Please associate at least one code/team project with this task, otherwise nobody can find this task when searching in the corresponding project(s). Thanks.

As a heads up @Tbayer we will need help from you this quarter in analysing any changes we make and help building the schemas that will help us evaluate that.

OK cool, looking forward to dedicating some time to this (let's talk more in our meeting this week). As we have started doing recently with other schemas, let's make sure there is an opportunity to vet the changes from a data analysis perspective to make sure the data will be suitable for answering the questions we have.

BTW, according to https://phabricator.wikimedia.org/T123634#1935198 this is already by far the largest EventLogging table in existence, at >500GB. So we should pay some attention to the sampling ratio and the expected event rate when we will discuss the details of the instrumentation.

I have no idea why you want to track users so much and I don't want to have anything to do with this endeavour, hence removing myself.

We should certainly revisit a lot of our EventLogging and determine whether we are using them or not. The MobileWebUIClickTracking for example is so big it is unusable right now.

@Jdlrobson, not yet, although it's on the list of potential things. The modal design changes don't themselves require this particular schema to be updated just yet, and I think we can probably redefine this schema in a manner where we capture which bucketing the user is in (if any) so that we may actually be able to defer a bit. This said, having data on the other tap targets is in and of itself for understanding relative tapthrough rates anyway.