Fix Security features tables cross-joining `ci_builds` to map pipeline -> security scans

From !62092 (closed) we have seen that VulnerabilityCountingService needs to join between non ci_* tables and ci_* tables.

But this will not be possible once we move ci_builds to a separate database.

The particular query comes from Pipeline#security_findings:

SELECT "security_findings".* FROM "security_findings"
INNER JOIN "security_scans" ON "security_findings"."scan_id" = "security_scans"."id"
INNER JOIN "ci_builds" ON "security_scans"."build_id" = "ci_builds"."id"
WHERE "ci_builds"."type" = 'Ci::Build' AND "ci_builds"."commit_id" = 5

This also affects Pipeline#has_security_findings? which is used in other places.

Options

1. De-normalize pipeline_id onto security_scans table
2. Change/refactor this feature such that we do not need these joins
3. Introduce another de-normalized join table that contains the columns we need to join without going to the ci_* table

Sharding team recommendation

Add a new column security_scans.pipeline_id:

And then you will update:

module EE
  module Ci
    module Pipeline

      prepended do
        has_many :security_scans, class_name: 'Security::Scan'

Such that the new query will be:

SELECT "security_findings".* FROM "security_findings"
INNER JOIN "security_scans" ON "security_findings"."scan_id" = "security_scans"."id"
WHERE "security_scans"."pipeline_id" = 5

This will be more performant and will remove the joins into ci_builds.

This will be similar to #333414 (closed) and can follow the similarly linked backfill approach of !37713 (merged) . It's also possible we could backfill both columns in 1 go.

Implementation plan:

• Migration for adding new column
• Migration for adding index and foreign key
• Background migration for backfilling data from ci_builds to security_scans
• Make sure we are inserting data for new column on save