Administration

Saved connections

Save source and target connection details once, reuse them to prefill new pipelines, and edit or delete them.

The Connections page is your connection repository: named, reusable connection details that prefill the source or target of a new pipeline. You never need one to create a pipeline. They save retyping the same host, database and user.

Open it from Connections in the top navigation. Each account has its own repository.

What a saved connection holds

StoredNot stored
Engine and nameA password, ever
Host, port, database and user
Engine-specific options, such as the capture mode of a source, the SQL Server authentication method or the Encryption setting
A secret reference, if you give one

A secret reference isn't a secret. It only names where the agent looks for the password on its own host, so SQL Flow stores it. You enter a password again each time a pipeline uses a saved connection, unless the connection names a secret reference. A pipeline keeps its own copy of its connection details, so changing or deleting a saved connection never affects an existing pipeline.

Each card shows the connection's address, sign-in and Encryption setting. Prefer shows in amber. Retired option, edit and save in red means the connection uses a transport option the agent no longer accepts: edit the connection and save it. Azure SQL and SQL database in Fabric connections have no Encryption row, because they always validate the server certificate.

Add a connection

  1. Open Connections and select Add connection.
  2. Enter a Name, for example orders-production. SQL Flow suggests one from the host and database.
  3. Choose the Engine.
  4. Fill in the fields the engine needs. They are the same fields the pipeline wizard shows for that engine, without a password. To use a secret reference, choose the Secret location (an environment variable or a file on the agent host) and give the Secret name. Both are optional.
  5. For SQL Server, PostgreSQL, MySQL, Oracle and Sybase ASE, choose the Encryption setting: Require (default) encrypts the connection and validates the server certificate; Prefer doesn't validate it, and for some engines connects unencrypted when the server offers no TLS. The form warns when you choose Prefer. See Connections commands for what each setting means per engine.
  6. Select Save.

A blank port takes the engine's default. A refused save shows its reason inside the dialog, which stays open.

A connection is identified by its engine and name. If you add a connection with the engine and name of an existing one, the dialog says A saved connection named "<name>" already exists. Saving replaces it. and the button reads Overwrite. Names that differ only in capitalization or punctuation count as the same name. The CLI's connection add refuses a name that exists instead.

Save a connection from the pipeline wizard

On a connection step of the wizard, select Save to repository. Enter a Connection name and select Save, or Overwrite if the name is already used for that engine. The wizard confirms with Saved <name>. Passwords are not saved.

Use a saved connection

On a connection step of the wizard, choose a connection under Saved connection (Select a saved connection…). It fills in the host, port, database, user and options, and the capture mode of a source. If it holds a secret reference, the reference is filled in too. Otherwise the wizard asks for the password: Passwords aren't saved to the repository: enter the password below for this connection.

With the CLI, pipeline create --source <connection> and --target <connection> take saved connections by name or ID. See Connections commands.

Edit a connection

  1. Select Edit on the connection.
  2. Change the fields. The Name and Engine are locked, because they identify the connection.
  3. Select Save.

If SQL Flow can't load all the details, it says Couldn't load all details for <name>. Retry editing before making changes. and the button reads Retry edit. Closing the dialog with unsaved changes asks Discard this connection?

Delete a connection

Select Delete and confirm Delete connection "<name>"? Pipelines already using it keep their own copy; new pipelines can no longer prefill from it. You can also delete one from the Saved connection list in the wizard.

Who can change saved connections

Every member of the account can see them. Adding, editing and deleting need the User or Account admin role. See Accounts and team.

See also

Previous
Projects

If this was useful,
our newsletter will be too.

Get monthly insights on data engineering, AI, and building critical infrastructure - direct from the Spectral Core team and CEO Damir Bulic.