PostgreSQL Database Agent
Overview
A PostgreSQL Database agent encapsulates the connection details for a single PostgreSQL database.
Usage
The PostgreSQL Database agent is used by the following connector:
Configuration
In addition to the common configuration shared by all agents, a PostgreSQL Database agent has the following settings:
| Name | Supports References | Description |
|---|---|---|
| Connection String | yes | The PostgreSQL connection string. |
| Enable Transactions | yes | Whether the TransactionScope ambient transaction context is used to manage transactions. |
| Operation Timeout | yes | The amount of time allowed for operations performed by the agent. |
Microsoft Entra authentication
The agent defaults to ConnectionString, so existing configurations continue to use the credentials in the connection string. The additional modes are DefaultCredential, WorkloadIdentity, ManagedIdentity, and ClientSecret.
For AKS, select WorkloadIdentity. The agent uses AZURE_TENANT_ID, AZURE_CLIENT_ID, and AZURE_FEDERATED_TOKEN_FILE by default; Tenant ID, Client ID, and Token File Path can be overridden in the agent configuration. Entra authentication still requires a connection string containing the PostgreSQL server, database, and database principal details, but it should not contain a password. Specify the database principal with Username; this is required for ClientSecret and recommended for the other Entra modes.
Authentication Mode is selected directly. Tenant ID, Client ID, Client Secret, and Token File Path can each be set directly or selected from the external configuration store in the agent editor.
The agent reuses one NpgsqlDataSource and one Azure credential. Existing pooled physical sessions remain authenticated, while every new physical connection obtains a valid token through the data source password provider. This avoids both per-query token requests and stale tokens in the connection pool.
See Microsoft Entra authentication for the mode and property descriptions.