​
Overview

dba.ai integrates with PostgreSQL to help you:

  • Monitor and optimize your PostgreSQL database performance
  • Detect and diagnose issues automatically
  • Receive AI-powered recommendations for improvements
  • Implement fixes with minimal effort

​
Connection Methods

You can connect dba.ai to your AWS RDS instances using one of two methods:

This method creates a role in your AWS Account that the dba.ai service is able to assume. It also creates all the necessary AWS resources to make cross account access secure. dba.ai will use a Cloudformation to create the necessary resources to connect to your AWS account and discover Postgres instances. You can view the Cloudformation template here.

Requirements:

  • AWS Secrets Manager enabled for your Postgres instance
  • an AWS role with select permissions

In order to do this, you will need to have access to an AWS role with the following permissions:

s3:CreateBucket - To create the LambdaZipsBucket
s3:PutBucketVersioning - For enabling versioning on the LambdaZipsBucket
s3:PutEncryptionConfiguration - For setting up bucket encryption
s3:PutBucketPublicAccessBlock - To configure the public access block settings
s3:GetObject - To retrieve lambda code from the source bucket
s3:PutObject - To copy objects to the destination bucket
s3:DeleteObject - For cleanup during stack operations

iam:CreateRole - To create the various IAM roles (CopyZipsRole, CrossAccountRole, LambdaExRoleForDBAI, RegistrationFunctionRole)
iam:PutRolePolicy - To attach inline policies to the roles
iam:AttachRolePolicy - To attach managed policies to the roles
iam:PassRole - Required for passing IAM roles to services like Lambda
iam:GetRole - To check if roles exist during resource creation or updates
iam:DeleteRole - For stack deletion operations
iam:DeleteRolePolicy - For removing policies during stack deletion
iam:DetachRolePolicy - For detaching managed policies during stack deletion

lambda:CreateFunction - To create Lambda functions (CopyZipsFunction, RegistrationFunction)
lambda:DeleteFunction - For stack deletion operations
lambda:UpdateFunctionCode - For stack updates
lambda:UpdateFunctionConfiguration - For stack updates
lambda:GetFunction - To check if functions exist during creation or updates
lambda:InvokeFunction - To invoke the custom resource functions

cloudformation:CreateStack - To create the stack initially
cloudformation:DescribeStacks - To check stack status
cloudformation:DescribeStackResource - To check resource status
cloudformation:DescribeStackResources - To get information about stack resources
cloudformation:GetTemplate - To retrieve the template
cloudformation:DeleteStack - To delete the stack if needed
cloudformation:UpdateStack - To update the stack if needed

sts:AssumeRole - When cross-account operations are performed

organizations:DescribeAccount - For organization-related operations (referenced in policies)
organizations:DescribeOrganization - For organization-related operations
organizations:ListAccounts - For organization-related operations
organizations:ListTagsForResource - For organization-related operations
ec2:CreateNetworkInterface - For Lambda network configurations
ec2:DescribeNetworkInterfaces - For Lambda network configurations
ec2:DeleteNetworkInterface - For Lambda network configurations
ec2:DescribeSecurityGroups - For security group verification
ec2:DescribeSubnets - For subnet verification
ec2:DescribeVpcs - For VPC verification
  1. Open a browser and login to your AWS account using a role with the permissions listed above.
  2. In the same browser navigate to dba.ai and login. Click on your user icon (lower left corner), then navigate to Settings > Connection -> Amazon Web Services -> Connect
  3. Click Open AWS Authorization Page in the popup. This will take you to the Cloudformation page within your AWS account
  4. Accept the Capabilities then click Create Stack and wait for the account to synchronize.

​
2. Direct Database Connection

A simpler method is to provide your database connection url directly to dba.ai. It is recommended to create a new role for this.

-- Create dedicated user for monitoring
CREATE USER dba_ai_monitor WITH PASSWORD 'strong-password-here';

GRANT pg_monitor TO dba_ai_monitor;
GRANT CONNECT ON DATABASE <your_database> TO dba_ai_monitor;

-- add privileges to public and any other schemas that you want dba to be able to read
GRANT USAGE ON SCHEMA public TO dba_ai_monitor;
GRANT SELECT ON ALL TABLES IN SCHEMA public TO dba_ai_monitor;
  1. Login with your browser at dba.ai, then click on your user icon (lower left corner) and navigate to Settings > Connections -> Amazon Web Services -> Connect
  2. Give the instance a name and provide the connection string using the user and password created above.
  3. Click Connect Database

​
Parameters and Configuration

For optimal performance with dba.ai, we recommend the following PostgreSQL parameter settings:

  • shared_preload_libraries: Include pg_stat_statements
  • track_activity_query_size: Set to at least 2048
  • pg_stat_statements.track: Set to ALL
  • log_min_duration_statement: Set to 1000 (milliseconds) or lower

​
Troubleshooting

If you encounter connection issues:

  1. Check that the database user has sufficient permissions
  2. Ensure your RDS instance is in the “Available” state

For persistent issues, please contact support@dba.ai with:

  • RDS instance identifier
  • Error messages received
  • Connection method used

​
Next Steps

After connecting your AWS RDS instance:

Was this page helpful?