Skip to main content

​
VegaFlow connections

A connection is a named, workspace-scoped configuration for a connector. Connector discovery tells the UI which properties to collect; connection validation checks whether the submitted configuration can be used.

​
Database connectors

Database connectors connect VegaFlow to structured data stores.

​
Application connectors

Application connectors connect VegaFlow to SaaS platforms and developer tools. These categories follow the VegaFlow connector registry: databases for PostgreSQL and applications for Airtable and GitHub. The registry can retain property schemas for future connectors without returning them from the enabled connector list. A schema asset is not an availability promise: saved connections currently accept only connector_id: postgres with connector_type: postgres.

​
Create a connection

  1. Open VegaFlow → Connections and select the connector.
  2. Enter a unique, descriptive name (up to the service limit) and optional description.
  3. Provide the fields from the connector property schema.
  4. Select an existing secret reference or create a vault and secret inline when permitted.
  5. Validate the connection.
  6. Save it, then grant only the collection or instance permissions consumers need.
The create and update models require a name, connector ID, connector type, and configuration object. Connector ID/type mismatches and disabled connector IDs are rejected.

​
Credential handling

Use secret references for passwords and private keys. A connection definition may safely retain the reference identifier; it must not copy the secret value into its description, run output, screenshots, or logs.

​
Catalog discovery

VegaFlow can retrieve a QuickFlow catalog from a saved connection. Use it to select valid source streams, tables, schemas, or columns instead of typing identifiers blindly.

​
Edit or delete

Revalidate after changing network, database, connector, or credential settings. Before deletion, identify QuickFlows that use the connection. Deletion and visibility remain permission-scoped.

​
Validation checklist

  • Confirm organization and workspace scope.
  • Confirm postgres is used for both connector ID and type.
  • Check host, port, database, and network reachability.
  • Check the selected secret reference and its permission.
  • Check whether the name conflicts with another active connection.