Connecting to Amazon Redshift ​

Connect Kyomi to your Amazon Redshift data warehouse for AI-powered analytics.

Connection Details ​

FieldDescriptionDefault
HostRedshift cluster endpointRequired
PortRedshift port5439
DatabaseDatabase to connect toRequired
Default SchemaDefault schema for queriespublic

Setup Steps ​

Step 1: Get Cluster Endpoint ​

  1. Open the Amazon Redshift console
  2. Select your cluster
  3. Find the Endpoint in the cluster details (e.g., mycluster.xxxx.us-east-1.redshift.amazonaws.com)

Step 2: Configure Connection ​

  1. In the datasource modal, select Amazon Redshift as the datasource type
  2. Enter your cluster Host (the endpoint without the port)
  3. Enter the Port (default: 5439)
  4. Click Connect to test the connection

Step 3: Select Database and Schema ​

  1. Choose your Database from the dropdown (typically dev or your custom database)
  2. Select a Default Schema (usually public)

Step 4: Configure Credentials ​

Enter your Redshift Username and Password.

Shared vs Personal Credentials

Shared Credentials: Admin configures once, all workspace users share the same database access.

Personal Credentials: Each user provides their own username/password for individual access control and audit trails.

Step 5: Configure Catalog ​

Select which schemas Kyomi should index:

  • Tables and columns from these schemas will appear in the catalog
  • The AI will use this information to help write queries
  • Leave empty to index all accessible schemas

SSH Tunnel (Optional) ​

For Redshift clusters in a private VPC, Kyomi supports SSH tunnel connections.

Configuration ​

  1. Enable Use SSH Tunnel in the connection settings
  2. Enter SSH connection details for a bastion host in your VPC:
    • SSH Host: Bastion server hostname
    • SSH Port: SSH port (default: 22)
    • SSH Username: Your SSH username
  3. The database connection will be tunneled through SSH

Required Permissions ​

Create a read-only user for Kyomi:

sql
-- Create user
CREATE USER kyomi_user PASSWORD 'SecurePassword123!';

-- Grant usage on schemas
GRANT USAGE ON SCHEMA public TO kyomi_user;
GRANT USAGE ON SCHEMA analytics TO kyomi_user;

-- Grant read access to all tables in schemas
GRANT SELECT ON ALL TABLES IN SCHEMA public TO kyomi_user;
GRANT SELECT ON ALL TABLES IN SCHEMA analytics TO kyomi_user;

-- Grant on future tables
ALTER DEFAULT PRIVILEGES IN SCHEMA public
  GRANT SELECT ON TABLES TO kyomi_user;
ALTER DEFAULT PRIVILEGES IN SCHEMA analytics
  GRANT SELECT ON TABLES TO kyomi_user;

Network Configuration ​

VPC Security Group ​

Ensure your Redshift cluster's security group allows inbound connections:

  1. Go to EC2 → Security Groups
  2. Find the security group attached to your Redshift cluster
  3. Add an inbound rule:
    • Type: Custom TCP
    • Port: 5439
    • Source: Kyomi's IP addresses (or your VPC CIDR for SSH tunnel)

Publicly Accessible ​

For direct connections (no SSH tunnel):

  1. Ensure your cluster is Publicly Accessible
  2. Or use VPC peering / PrivateLink

Troubleshooting ​

"Connection refused" or "Connection timed out" ​

  • Verify the endpoint and port are correct
  • Check security group allows inbound connections on port 5439
  • Verify the cluster is running (not paused)
  • Check if cluster is publicly accessible (or use SSH tunnel)

"Invalid username or password" ​

  • Verify credentials are correct
  • Note: Redshift usernames are case-sensitive

"Permission denied" when querying ​

  • Verify the user has SELECT privilege on the tables
  • Check schema permissions (USAGE on schema)

Can't see expected tables ​

  • Verify the user has access to the schema
  • Check "Schemas to Index" includes the desired schemas
  • External schemas (Spectrum) may require additional permissions

Redshift Serverless ​

For Redshift Serverless:

  1. Find the Workgroup endpoint in the Redshift Serverless console
  2. Use the endpoint as the Host
  3. Port is still 5439
  4. Authentication works the same as provisioned clusters

Redshift Spectrum ​

If you use Redshift Spectrum for querying S3 data:

  • External schemas will appear in the catalog if indexed
  • Ensure the user has access to the external schema
  • Spectrum queries may take longer due to S3 access
sql
-- Grant access to external schema
GRANT USAGE ON SCHEMA spectrum_schema TO kyomi_user;
GRANT SELECT ON ALL TABLES IN SCHEMA spectrum_schema TO kyomi_user;

Performance Tips ​

  • Distribution keys: Queries joining on distribution keys are faster
  • Sort keys: Filtering on sort keys improves performance
  • VACUUM/ANALYZE: Run regularly for optimal query planning
  • Concurrency scaling: Enable for handling multiple concurrent queries

Additional Resources ​

← Back to Datasources | Back to Docs