Track HiveGrid Cloud planning reference

This commit is contained in:
david committed 2026-06-26 20:48:01 +02:00
1 parent b3d5ea6764
commit e341c5938a
5784 files changed
+1104929 -3

No files matched your search

@@ -0,0 +1,22 @@
-- 04/vector-store: enable the pgvector extension on the application database.
--
-- This is the foundation for `packages/vector-store`'s pgvector adapter. Each
-- collection lives in its own table named `vec_<namespace>_<collection>` with
-- a `vector(<dim>)` column plus HNSW + GIN + tenant-id indexes. The tables are
-- created on demand at runtime by the adapter (`createCollection`); only the
-- extension itself is provisioned here so the rest of the migration history
-- stays free of dynamic vector-table churn.
--
-- Notes for operators:
-- * `CREATE EXTENSION` requires superuser privileges. On managed Postgres
-- (RDS, Supabase, Neon, …) the extension is typically pre-allowlisted and
-- this statement is idempotent.
-- * Self-hosted Postgres needs the pgvector binaries installed
-- (`apt install postgresql-16-pgvector` or the `pgvector/pgvector:pg16`
-- Docker image used in CI). If the binaries are missing, the migration
-- fails fast — preferable to silent breakage when the first `createCollection`
-- call runs.
-- * Rollback drops the extension via `DROP EXTENSION IF EXISTS vector CASCADE`;
-- CASCADE destroys every `vec_*` table, so back up before reverting.
CREATE EXTENSION IF NOT EXISTS vector;