Skip to content

Connections

A connection represents a set of credentials and configuration for reaching an external data source. DataRecs supports relational databases (PostgreSQL, MySQL, Oracle, IBM DB2, Microsoft SQL Server), cloud data warehouses (Snowflake), Google BigQuery, Google Sheets, Excel on SharePoint, and BI sources (Tableau, Looker).

  • Credentials are envelope-encrypted and stored in OpenBao — never in plaintext.
  • Network access to your data source is tightly controlled: only authorised egress traffic leaves the platform.
  • Each connection is validated at creation time before its credentials are stored.
SourceIdentifier
PostgreSQLpostgres
MySQLmysql
IBM DB2db2
Oracleoracle
Microsoft SQL Servermssql
Snowflakesnowflake
Google BigQuerybigquery
Google Sheetsgoogle_sheets
Excel on SharePointexcel_sharepoint
Tableautableau
Lookerlooker

Most relational sources authenticate with a username and password. Snowflake uses RSA key-pair authentication instead — see Connecting to Snowflake for the setup.

  1. Create — provide the source type and its credentials (for databases: host, port, username, password, database).
  2. Test — validate connectivity (or the impersonation grant, for keyless types) via connections test, without creating anything.
  3. Use — reference the connection in reconciliation jobs.
  4. Rotate — update credentials at any time; active jobs will pick up the change on their next run.

A connection has one owning workspace (controls who can modify it and its sharing) and can be shared into additional workspaces so more than one team can reference it in a job.