Description of Documentation Need

The README's "Notes" section (from line 27) documents a provider-bug workaround: manually delete a peered network, release an allocated IP, and terraform state rm three resources including google_compute_global_address.postgres and google_service_networking_connection.postgres. Neither managed resource exists anywhere in the current main.tf (0 matches) — the blueprint was refactored to consume the shared cloudsql-instance module, which only references an existing PSA connection via psa_config.private_network and never creates the address reservation or the peering. Anyone following the documented teardown steps references resources Terraform has never had in state, and the state rm commands fail immediately. Additionally, the refactor means PSA peering must pre-exist on the target VPC before apply — but the "Requirements" section only says "An existing VPC."

Target Audience

Operators deploying or tearing down the blueprint; anyone troubleshooting a failed apply.

Proposed Location

blueprints/il5/postgresql/README.md — remove/replace the stale "Notes" section; extend "Requirements."

Content Outline / Draft

1. Remove the stale provider-bug cleanup procedure (resources no longer managed).
2. Add to "Requirements": an existing VPC with Private Service Access peering already configured on network_name / network_project_id.
3. (Optional) A short note on what the blueprint does not create (PSA address reservation, peering connection), so teardown expectations match reality.

Compliance Context (if applicable)

None specific — accuracy/teardown-safety fix. Affects both FRH and IL5 consumers (blueprints/fedramp-high/postgresql is a symlink to this path).