Amazon Redshift
Connect an Amazon Redshift cluster so Qubit can read the tables in one schema. You can copy the data into Qubit on a schedule (Sync Mode) or query the cluster live (Direct Mode). Clusters inside a private network can be reached through an SSH bastion host.
Before you start
- Create a dedicated read-only user for Qubit and grant it
SELECTon the tables in the schema you want to use, plusUSAGEon the schema. - Note the cluster endpoint (host), port, database, schema, and AWS Region.
- If the cluster is not publicly accessible, have an SSH bastion host in the same network, with a user and private key Qubit can use.
Connect Redshift
- Open Data Catalog from the sidebar, click Add Data, then choose Connect a database.
- On the New Data Source page, click Redshift. It is under Databases.
- Enter a Connection Name.
- Fill in the connection details:
- Host: the cluster endpoint, without the port and database.
- Port: defaults to
5439. - Database: the database name.
- Schema: the schema to read. This field is required.
- User and Password: the read-only user you created.
- AWS Region: the region of the cluster, for example
me-south-1.
- Under Connection Mode, choose Sync Mode to copy the data into Qubit, or Direct Mode to query the cluster live.
- To reach the cluster through a bastion host, turn on Connect via SSH tunnel and enter the SSH host, SSH port, SSH username, SSH private key (PEM), and the key passphrase if it has one. The tunnel always uses Direct Mode, so no data is copied into Qubit.
- For Sync Mode, choose a Sync Schedule. The default is Daily.
- Click Create.
Qubit tests the connection before saving it, then takes you to the Data Catalog. You cannot change the connection mode or the SSH tunnel after the connection is created.
If the test fails
The form shows the error message:
- "Authentication failed": check the user and password.
- "Could not reach" or "timed out": check the endpoint and port, and that the cluster's security group accepts connections from Qubit. For a private cluster, use the SSH tunnel.
- "Schema does not exist": check the schema name.