ci: batch pg/integration test jobs + faster container health-checks

[pyramation]

Summary

Follow-up to the runner-downsizing PR. Measured on a recent run-tests.yaml main run, each pg/integration/ai matrix job spends only ~6s in its actual Test step but ~48s in fixed overhead — "Initialize containers" alone is ~25s/job (47%). So we were paying per-job service-container + install overhead ~30 times over.

Two changes:

1. Batch the per-package matrix jobs, mirroring the existing unit-tests tier (one job runs several packages' tests sequentially against one shared service container):

   • pg-tests: 19 jobs → 3 (pg-core, pg-postgres, pg-graphql)
   • integration-tests: 10 jobs → 2 (integration-graphql, integration-uploads)
   • Per-package env (TEST_DATABASE_URL, BUCKET_NAME) is hoisted to the job-level env — harmless for packages that ignore it.

2. Tighten service-container health-checks --health-interval 10s → 3s (retries 5 → 10) for Postgres/MinIO across all three tiers. A container that's ready in ~2s previously wasn't marked healthy until the first 10s poll, wasting ~7s/job.

Net: ~29 DB jobs → ~5, removing ~24× the ~48s fixed overhead (plus the health-check savings on what remains). Test coverage is unchanged — same packages, just grouped. unit-tests (no services) and build are untouched.

Link to Devin session: https://app.devin.ai/sessions/180599d6e0c1419994599efb1eabd9b5
Requested by: @pyramation

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring