Skip to content
DataMaker Docs

Connections

A connection is a configured target system that DataMaker can push generated data into — or pull existing records out of. Once configured, every template can be “exported” to that connection from the UI, the API, a scenario, or the agent.

DataMaker supports three families of connections:

  • Databases — Postgres, MySQL, MongoDB, MSSQL, Oracle, IBM DB2.
  • REST endpoints — any HTTP API, with custom headers and auth.
  • SAP OData — V2 and V4 services, with auto-CSRF and $metadata discovery.

Anatomy of a connection

{
  "id":      "conn_abc123",
  "name":    "S/4 Sandbox",
  "type":    "sap_odata",
  "url":     "https://my-sap.example.com/sap/opu/odata/sap/API_BUSINESS_PARTNER_SRV",
  "auth":    { "type": "basic", "secret_ref": "sec_xyz" },
  "metadata": { "entitySets": ["A_BusinessPartner", "A_BPAddress", ...] },
  "lastVerifiedAt": "2026-04-25T18:14:00Z"
}

Auth secrets are stored encrypted; you reference them by ID, never by value, in scenarios and the API.

Verifying a connection

When you create a connection, DataMaker runs a quick verification:

  • Databases: opens a connection, runs SELECT 1 (or equivalent).
  • REST: makes an authenticated OPTIONS or GET to the base URL.
  • SAP OData: fetches $metadata and parses the entity sets.

If verification fails, the connection is created in an unverified state. You can still edit and retry. See Troubleshooting → Common errors if your auth flow needs a one-off CSRF setup or proxy.

Mapping a template to a connection target

Templates and connections are independent. To push, DataMaker needs to know which template field maps to which target column / property / entity property.

The first time you push, the UI shows a mapping screen that auto-suggests by name match. You can:

  • Override individual mappings.
  • Skip target columns (DataMaker leaves them at their default / null).
  • Save the mapping so subsequent pushes are one-click.

Mappings are stored on the template per-connection.

Push, fetch, both

Most teams push more often than they fetch — the workflow is “generate synthetic data, write it into the test system”. But for SAP regression workflows, you also fetch existing records into a saved set you can run regression against. See Workflows → SAP regression.

The same connection is used for both directions; nothing extra to configure.