pal-e-docs

Validation: pal-e-api#252 — Partial indexes for mermaid blocks and architecture notes

validation-252-2026-04-12 Doc

validation partial
Project

Ticket

forgejo_admin/pal-e-api#252 — Add partial indexes for mermaid blocks and architecture notes
Board item: #944 on board-pal-e-docs
PR: #253 (merged)
Shipped: Two partial indexes — ix_blocks_block_type_mermaid on blocks table and ix_notes_note_type_architecture on notes table — to speed up filtered queries.

Environment

Prod cluster (archbox), namespace pal-e-docs. Database: paledocs on pal-e-postgres-1 in namespace postgres.

Checks

# Criterion How to Verify Result Evidence
1 PR #253 merged to main Forgejo PR list, state=closed PASS PR #253 state=closed, merged=true
2 Pod running in pal-e-docs namespace <code>kubectl get pods -n pal-e-docs</code> PASS pal-e-docs-76b5f66c69-qxlg6 Running, 0 restarts, 14d age
3 Index <code>ix_blocks_block_type_mermaid</code> exists on blocks table <code>kubectl -n postgres exec pal-e-postgres-1 -- psql -U postgres -d paledocs -c "\di"</code> FAIL Index not present in <code>\di</code> output. 27 indexes listed, neither partial index exists.
4 Index <code>ix_notes_note_type_architecture</code> exists on notes table <code>kubectl -n postgres exec pal-e-postgres-1 -- psql -U postgres -d paledocs -c "\di"</code> FAIL Index not present. Current alembic_version: <code>s9n0o1p2q3r4</code>
5 EXPLAIN uses mermaid partial index Blocked — index does not exist BLOCKED N/A

Verdict

PARTIAL — PR merged successfully, pod is healthy, but the alembic migration has not been applied to production. The two partial indexes (ix_blocks_block_type_mermaid, ix_notes_note_type_architecture) do not exist yet. Needs alembic upgrade head run against prod database.

Discovered Issues

  • Pod image is 14 days old (tag 89f663bb...). The deployment may not have been updated after PR #253 merged. Verify CI built and pushed the new image, and that ArgoCD/kustomize picked up the new tag.
  • Migration alembic upgrade head needs to be run on prod — either via a job or manual exec into the pod.