Self-hosted Logflare with BigQuery backend cannot work out of the box

# Bug report <!-- ⚠️ We receive a lot of bug reports which have already been solved or discussed. If you are looking for help, please try these first: - Docs: https://docs.supabase.com - Discussions: https://github.com/supabase/supabase/discussions - Discord: https://discord.supabase.com Before opening a bug report, please verify the following: --> - [x] I confirm this is a bug with Supabase, not with my own application. - [x] I confirm I have searched the [Docs](https://docs.supabase.com), GitHub [Discussions](https://github.com/supabase/supabase/discussions), and [Discord](https://discord.supabase.com). ## Describe the bug I used Claude to help me play "spot the difference" with some of this, so forgive me for also using it to help with a writeup. This is likely both a documentation and code bug. The first problem is that when starting up Logflare with a BigQuery backend, Logflare creates the necessary tables, but does not create the necessary columns within those schemas up-front. The columns will get created as rows enter BigQuery, which is fine, but if I don't use several products those tables will remain with a default, incomplete set of columns causing errors not only on that tab, but depending on the table *all of the tabs* because of details in the below writeup. There is also a disparity between what Vector sends for columns vs what Studio queries against in some of the Logs tabs, meaning even if the rows and columns get created eventually some of those tabs will *never* work without additional work. Claude-generated writeup/additional details follows: When using BigQuery as the Logflare backend for self-hosted Supabase, all Studio log tabs fail with invalidQuery errors. There are three issues: 1. Vector sends fewer fields than Studio queries expect The self-hosted Vector pipeline sends a limited set of metadata fields, but Studio's queries reference fields that only exist on the Supabase Platform. BigQuery auto-creates table schemas from ingested data…