Daniel Thorn
Daniel Thorn
nothing misrepresented, but from the [same PR on another product](https://github.com/mozilla-rally/rally/pull/47#issuecomment-1164708798): The main benefit here for Glean products is that this CI job will allow people to see if changes to...
@isabelrios this was rebased, and checks are now passing.
yes, I can test it.
> What would probably break this idea is that for queries over all of main ping, we're introducing 100x the schema overhead. I'd be that's a real drag on query...
+1, let's make this issue for actually doing it
this matches my understanding of the situation. my concern is that for landfill the subscription is our only copy of data, so if the beam job were to crash for...
I'm inclined to say that we should instead provide an `official_builds` table and a UDF in BigQuery, and make this a post-processing thing.
that bq table may even be useful as a public table to people outside of mozilla, if we made a blog post about it being available.
~This should no longer be blocked, because we can now use the `*AsyncClient`s from upstream~ edit: upstream does not expose async clients
Cool. It looks like `AfterPane.elementCountAtLeast()` can be used to collect an exact number of records, and can be combined with an `AfterProcessingTime` trigger using [composite triggers](https://beam.apache.org/documentation/programming-guide/#composite-triggers) to additionally use a...