Follow-up from "Geo: Add models to prepare for project repo replication v2"

The following discussion from !209742 (merged) should be addressed:

• @c_fons started a discussion: (+2 comments)

  thought: There will be an ID column acting as PK in this table. Setting this up seems redundant - why not use the PK created by the migration instead? Besides, this might cause database-related errors. See this discussion: #578043 (comment 2838041372) and its associated MR: !210167 (merged)