I've been a bit sloppy in hard-coding project IDs into some of the scheduled jobs here (e.g., when writing into a particular BigQuery dataset). That makes this project hard to lift from one project to another (including when I'm doing some testing in a sandbox account).
Probably better to store default project IDs in an airflow variable and pick it up where appropriate. Variables could be stored in a plain text file and re-sync'd on deploy.
I've been a bit sloppy in hard-coding project IDs into some of the scheduled jobs here (e.g., when writing into a particular BigQuery dataset). That makes this project hard to lift from one project to another (including when I'm doing some testing in a sandbox account).
Probably better to store default project IDs in an airflow variable and pick it up where appropriate. Variables could be stored in a plain text file and re-sync'd on deploy.