Connector reference

The connector catalog is workspace- and region-aware. Always read it before rendering a connection form or assuming a connector is enabled. Browse connector guides by type for database and API/application configuration examples.

List enabled connectors

GET {api-base}/orgs/{org_id}/workspaces/{workspace_id}/vegaflow/connectors

Each item identifies the connector, display name, category, connector type, supported roles and presentation metadata. The catalog entry is the authority for availability—not an icon shown in a marketing page.

Read the property schema

GET {api-base}/orgs/{org_id}/workspaces/{workspace_id}/vegaflow/connectors/{connector_id}/properties

The response contains a JSON Schema and UI schema for the connection configuration supported by that connector release. Clients should validate against the returned schema and preserve unknown server-managed fields.

Capability vocabulary

A connector can expose only a subset. VegaFlow rejects a QuickFlow whose selected sync mode is absent from the discovered stream.

Configuration and secret fields

The property schema distinguishes ordinary configuration from secret references.

{
  "host": "orders.internal.example",
  "port": 5432,
  "database": "orders",
  "username": "vegalake_sync",
  "ssl_mode": "verify-full",
  "password_auth": {
    "password_ref": {
      "namespace": "production-data",
      "key": "orders-sync-password"
    }
  }
}

Store passwords and keys as secret references containing namespace and key. A saved definition can retain these identifiers; secret values must stay out of configuration examples, descriptions and logs. References are resolved only for authorized work at execution time.

Validate before saving

POST {api-base}/orgs/{org_id}/workspaces/{workspace_id}/vegaflow/connections/validate
Content-Type: application/json

Validation checks the submitted connector configuration from the selected network/compute context. Depending on the connector it can verify DNS, TLS, authentication, namespace access, discovery privileges and destination write prerequisites.

Validation is a point-in-time test. Firewall policy, credential expiry and source permissions can change before a later run.

Catalog discovery

GET {api-base}/orgs/{org_id}/workspaces/{workspace_id}/vegaflow/connections/{connection_id}/quickflow/catalog?cluster_id={cluster_id}

Discovery requires use access to both the connection and compute environment. It returns only namespaces and streams allowed by the connection configuration and source principal.

For a large source, restrict allowed databases/schemas before discovery. Do not ask each UI refresh to enumerate an entire enterprise server.

PostgreSQL connector

The production PostgreSQL connector supports connection testing, schema/table discovery and the sync modes advertised for each discovered stream. Saved connections require postgres for both connector ID and type. Use the PostgreSQL guide for database settings, SSL modes and password or SSH-tunnel authentication.

GitHub source and Airtable destination definitions remain preview. A workspace must list a connector as enabled before it can be used in a saved connection or runnable flow. See connector availability.

Configuration versioning

Updating a connection creates a new resource version. New runs resolve the active version; completed runs retain the version identity they used. A retry can be rejected when the new connection configuration is incompatible with the selected immutable flow version.