claude-memory/graph/decisions/object-handler-cloudsql-silver-layer-architecture-e80e16.md

id	type	title	tags	importance	confidence	created	updated	relations
e80e161a-4be4-428d-8590-267b9ab4cc7c	decision	Object-handler CloudSQL silver layer architecture	esb-monorepo object-handler architecture cloudsql decision	0.8	0.8	2025-12-10T19:55:26.314546+00:00	2026-03-04T01:48:02.005397+00:00	target type direction strength 6d3dcf69-9c0c-4d8a-b1a7-27267c07da51 BUILDS_ON outgoing 0.5 target type direction strength cd528dd3-8328-41c5-9652-81c410c0d344 BUILDS_ON incoming 0.5 target type direction strength edge_id 35b13826-2f86-460a-a087-c15ba92a9d3e RELATED_TO incoming 0.58 f335cae2-3e63-477e-a7ea-aaac59077089 target type direction strength edge_id 8a4ae0ed-5bac-459b-bdd8-d3075bfc108d RELATED_TO incoming 0.65 5718558e-60c9-487f-ab46-9dee64c3385c target type direction strength edge_id 43b222b2-abef-4ae0-8d3f-45c7a588eaa7 RELATED_TO incoming 0.68 0a395691-6b0f-4172-a4f6-a45057d131b0

For object-handler silver layer (CloudSQL) output: Use Cloud SQL Python Connector + pg8000 driver (not SQLAlchemy). Connection method chosen because: (1) Works in Docker locally AND Cloud Functions, (2) IAM auth eliminates password management, (3) No VPC connector needed. Postgres functions handle upsert logic, so raw SQL calls are sufficient - SQLAlchemy would be overkill. Data flow: raw_data stripped before silver layer, full data goes to Firestore bronze layer.