database-per-service

Querying Three Service Databases in Parallel to Build a Unified View

Each microservice owns its own database: users, orders, and products. Building a customer dashboard requires querying all three -- but doing it sequentially triples the latency. This workflow uses FORK_JOIN to query all three databases in parallel, then composes a unified view from the results.

Workflow

userId
  |
  v
  FORK_JOIN ---------------------------------+
  |                |                         |
  v                v                         v
+------------------+  +--------------------+  +----------------------+
| dps_query_       |  | dps_query_         |  | dps_query_           |
| user_db          |  | order_db           |  | product_db           |
+------------------+  +--------------------+  +----------------------+
  name: "Alice"        orderCount: 12          recentProducts:
  tier: "premium"      lastOrder: "ORD-999"    ["Widget", "Gadget"]
  |                |                         |
  +------ JOIN ---+--------------------------+
               |
               v
      +--------------------+
      | dps_compose_view   |   unified view from 3 databases
      +--------------------+

Workers

QueryUserDbWorker -- Queries the user database for userId. Returns name: "Alice", tier: "premium".

QueryOrderDbWorker -- Queries the order database. Returns orderCount: 12, lastOrder: "ORD-999".

QueryProductDbWorker -- Queries recently viewed products. Returns recentProducts: ["Widget", "Gadget"].

ComposeViewWorker -- Builds a unified view from all three database query results.

Tests

8 unit tests cover each database query and the view composition.

Running

See ../../RUNNING.md for setup and execution instructions.