improvement(db): disable embedding FTS fastupdate and index delete-action foreign keys

Inserts into embedding could stall for seconds merging the shared GIN pending
list in the foreground, while holding the document row lock. Turn fastupdate
off (the 0364 pattern) and drain the existing list once.

User and workspace deletion ran sequential scans for four referential actions
inside the deleting transaction, holding row locks other writers queued behind.
Add concurrent indexes backing document.uploaded_by,
user_table_rows.created_by, user_table_definitions.created_by and
execution_large_values.workspace_id.

Author: waleedlatif1

packages/db/migrations/0410_gin_fastupdate_fk_indexes.sql

@@ -0,0 +1,23 @@
+-- migration-safe: Metadata-only storage option; existing and new workers retain the same index and search semantics.
+ALTER INDEX "public"."emb_content_fts_idx" SET (fastupdate = off);--> statement-breakpoint
+-- Release the DDL lock before maintenance and index builds. Everything below is safe to replay after a partial migration.
+COMMIT;--> statement-breakpoint
+-- Disabling fastupdate does not flush existing pending entries. Drain them once without rebuilding the index.
+SET statement_timeout = '5min';--> statement-breakpoint
+SELECT pg_catalog.gin_clean_pending_list('"public"."emb_content_fts_idx"'::regclass);--> statement-breakpoint
+SET statement_timeout = 0;--> statement-breakpoint
+-- Indexes backing user and workspace delete actions, which otherwise scan these tables inside the deleting transaction.
+SET lock_timeout = 0;--> statement-breakpoint
+-- migration-safe: replay replaces only this new index to recover an interrupted concurrent build; existing indexes remain available.
+DROP INDEX CONCURRENTLY IF EXISTS "doc_uploaded_by_idx";--> statement-breakpoint
+CREATE INDEX CONCURRENTLY IF NOT EXISTS "doc_uploaded_by_idx" ON "document" USING btree ("uploaded_by") WHERE "document"."uploaded_by" IS NOT NULL;--> statement-breakpoint
+-- migration-safe: replay replaces only this new index to recover an interrupted concurrent build; existing indexes remain available.
+DROP INDEX CONCURRENTLY IF EXISTS "execution_large_values_workspace_id_idx";--> statement-breakpoint
+CREATE INDEX CONCURRENTLY IF NOT EXISTS "execution_large_values_workspace_id_idx" ON "execution_large_values" USING btree ("workspace_id");--> statement-breakpoint
+-- migration-safe: replay replaces only this new index to recover an interrupted concurrent build; existing indexes remain available.
+DROP INDEX CONCURRENTLY IF EXISTS "user_table_def_created_by_idx";--> statement-breakpoint
+CREATE INDEX CONCURRENTLY IF NOT EXISTS "user_table_def_created_by_idx" ON "user_table_definitions" USING btree ("created_by");--> statement-breakpoint
+-- migration-safe: replay replaces only this new index to recover an interrupted concurrent build; existing indexes remain available.
+DROP INDEX CONCURRENTLY IF EXISTS "user_table_rows_created_by_idx";--> statement-breakpoint
+CREATE INDEX CONCURRENTLY IF NOT EXISTS "user_table_rows_created_by_idx" ON "user_table_rows" USING btree ("created_by") WHERE "user_table_rows"."created_by" IS NOT NULL;--> statement-breakpoint
+SET lock_timeout = '5s';