Drop this run's test database when the job ends, instead of two days later

[mkreyman]

Same omission as home_care_billing#1138.

Nothing dropped these. infra-ci-db-reap swept them on a 2-day retention, which made a backstop into the primary mechanism — 440 live databases on the shared CI instance on 2026-08-12, each waiting up to two days for a sweep that only exists because this step did not. There was never a rationale for keeping them; it was an omission.

The reaper stays: a cancelled or killed job never reaches its own cleanup step, so end-of-job dropping cannot be the only mechanism. Backstop second, not first.

• One drop covers the run — Repo, AdminRepo and HeavyReadRepo all resolve to loopctl_test${MIX_TEST_PARTITION}.
• if: always() — a failed suite strands a database exactly like a passing one.
• Best-effort, so a cleanup problem cannot turn a green run red.
• PGPORT is already job-scoped to 5433, the CI-only instance.

The only existing DROP in this workflow is for the scale-test database, which runs on a hosted runner against a service container — never the one accumulating here.

Reviewed inline (this session cannot dispatch agents/workflows) per the CLAUDE.md gate clause.