Add organization_id column to bulk_import_exports

What does this MR do and why?

Part 1 of 2: Add organization_id column to bulk_import_exports

First step toward an organization_id sharding key. Schema-only, no data changes:

• Adds nullable organization_id (bigint)
• Adds a concurrent index on it
• Adds a FK to organizations (on_delete: :cascade, NOT VALID; validation comes in a later phase)

The backfill follows in Part 2.

Note for reviewers: this is a stacked MR. Please review Part 2 as well: !241089 (merged).

Migration details

db/migrate/20260617101501_add_organization_id_to_bulk_import_exports.rb
db/post_migrate/20260617101502_add_concurrent_index_organization_id_on_bulk_import_exports.rb
db/post_migrate/20260617101503_add_foreign_key_constraint_on_bulk_import_exports_organization_id.rb

MR acceptance checklist

• Migrations tested up and down locally
• Database review

References

• Issue #600457 (closed), epic &21584 (closed)
• Row counts (database-lab): bulk_import_exports ~7.6M, bulk_import_export_batches ~808K. Too large for an inline backfill, so the work uses the BBM path across ~%19.2–%19.4.
• Follows the sibling bulk_import_trackers rollout.