Release 2642
Trello card
Context
We want to send entity information to BigQuery with any personally identifying attributes anonymised. PII that is not useful for analysis (such as emails, phone numbers, etc) are not sent at all, along with various free text attributes that aren't easily aggregated.
Changes proposed in this pull request
- Whitelist attributes for BigQuery
Guidance to review
I opted to blacklist the delayed jobs table; we would need to anonymise the handler attribute as it may inadvertently end up with PII in it (if someone creates a job which accepts an email address, for example) and it's not useful without that.