Preservation and operations
Preserving the archive requires both the source bytes and the catalog that explains them. Serving it also creates derivatives and operational state that belong to a particular environment.
Cloud preservation explains canonical originals, served derivatives, database transfer packages, upload integrity, and recovery boundaries.
Jobs, ownership, and synchronization explains bounded background work and the different directions in which data moves. Core archive metadata moves from development to production. Life-era definitions and memberships originate in production. People and face review originate locally. Each environment keeps its own media attachments and derivatives.
The documentation site you are reading is separate from the Rails application: Jekyll builds static pages, and a Cloudflare assets-only Worker serves the output. The archive application has its own runtime, database, and workers.