Skip to main content

Data Architecture

Summary​

HR Suit uses PostgreSQL with a transitional ownership model: the monolith uses the default schema, extracted services use service-owned schemas, and the platform boundary manages shared tenant metadata and optional tenant database provisioning.

Audience​

  • Developers and QA engineers
  • DevOps engineers
  • Implementation partners
  • Solution architects

Overview​

The active monolith retains persistence in PostgreSQL’s default schema. Extracted services use separate EF Core contexts and service-owned schemas, creating logical ownership boundaries even where infrastructure is shared. Services apply EF Core migrations; the monolith still uses a transitional initialization mechanism.

The service-schema nodes are representative. They do not imply that only two service schemas exist. The optional database-per-tenant path is a platform capability and is not a universal storage model for every service.

Ownership and consistency​

  • The component that owns a schema is responsible for its data access and evolution.
  • Cross-service state is exchanged through APIs or integration events rather than direct ownership of another service’s domain tables.
  • Audit records capture selected business or administrative changes.
  • Inbox or processed-event records support idempotent event handling.
  • Transactional outbox records couple a domain change with an event awaiting publication.
  • Payroll and document-related processing maintain selected local read models from other domains’ events.

Tenant-scoped access is enforced through a combination of application and persistence mechanisms. Implementation varies by boundary; multi-tenancy describes this without asserting one universal filter.

Documentation boundary​

This page intentionally excludes table names, columns, indexes, connection configuration, migration commands, retention policy, and production topology. Those require later database and operations documentation with environment-specific review.

See Also​

Keywords​

  • Data ownership
  • Database migration
  • Idempotency

Revision Information​

  • Last reviewed: 2026-07-14
  • Owner: architecture-team
  • Status: Draft