Google Cloud SQL - PostgreSQL + Datagrid integration

Overview

What is Google Cloud SQL - PostgreSQL: Cloud SQL for PostgreSQL is Google Cloud's fully managed PostgreSQL service for relational databases on Google Cloud Platform, used as the transactional backbone for applications on Compute Engine, GKE, Cloud Run, and App Engine. Cloud SQL automates backups, failover, replication, encryption, and patching for PostgreSQL databases.

How to integrate Google Cloud SQL - PostgreSQL with Datagrid

Datagrid connects to Cloud SQL for PostgreSQL through a database connection, imports selected database objects, and makes that data available for agent execution. Data flows one way, from Cloud SQL into Datagrid; agents read the synced records, cross-check them against other connected systems, answer natural language questions, and generate reports. The same database connection pattern also applies to the Datagrid's generic PostgreSQL integration. Set up the integration by preparing Cloud SQL access, connecting the instance, authenticating with PostgreSQL credentials, and configuring recurring data syncs.

Prepare Cloud SQL access

Start by confirming that Datagrid can reach the instance and that the database account has the least access required for the workflows agents will run. For direct public IP connections, Cloud SQL requires the client IP in the instance's authorized networks list, configured under the instance's Connections tab in the Google Cloud console.

Use this access-planning block before changing the instance connection settings:

authorized_networks_example:

cloud_sql_section: Connections

instance: project-db-instance

network_name: datagrid-public-ip

client_ip: DATAGRID_CLIENT_IP

purpose: direct public IP connection

Connect the instance in Datagrid

Follow these steps in Datagrid after the Cloud SQL instance is reachable.

1. In Datagrid, click Create, then select Connect Apps.

2. Search for Google Cloud SQL - PostgreSQL in the integration list.

3. Enter your server name, database name, username, and password.

4. Select the data objects you want to import.

5. Click Start First Import.

Use this example as a connection-planning artifact before entering values in Datagrid:

connection_example:

integration: Google Cloud SQL - PostgreSQL

host: project-db-instance.example-postgres.googleusercontent.com

port: 5432

database: operations

username: datagrid_readonly

password: example-password-stored-in-datagrid

ssl_mode: require

sync_example:

selected_objects:

- public.orders

- public.invoices

- reporting.weekly_project_status

object_types:

- tables

- views

- stored procedures

schedule: recurring

Authenticate with PostgreSQL credentials

Create a dedicated database user for Datagrid before starting the first import. The integration uses credential-based authentication: host, port, database name, username, and password, mapping to Cloud SQL's built-in database authentication. Limit the user's read access to the schemas agents need.

Use grants like this for table and view access in the selected schemas:

CREATE USER datagrid_readonly WITH PASSWORD 'example-password-stored-in-datagrid';

GRANT CONNECT ON DATABASE operations TO datagrid_readonly;

GRANT USAGE ON SCHEMA public, reporting TO datagrid_readonly;

GRANT SELECT ON ALL TABLES IN SCHEMA public, reporting TO datagrid_readonly;

Configure recurring data syncs

Use the sync settings to decide which operational records agents can read and how often Datagrid refreshes them. Datagrid can sync tables, views, and stored procedures, reading from Cloud SQL for PostgreSQL into Datagrid only, on configurable recurring sync schedules.

Use this sync configuration block to document the import scope before the first run:

recurring_sync_configuration:

integration: Google Cloud SQL - PostgreSQL

selected_schemas:

- public

- reporting

selected_objects:

- public.orders

- public.invoices

- reporting.weekly_project_status

refresh_schedule: recurring

Review Cloud SQL's connection options if your instance access model changes before the first import.

Why use Google Cloud SQL - PostgreSQL with Datagrid

Connect Cloud SQL for PostgreSQL to Datagrid when operators need answers and action from transactional data without pushing every request through a separate reporting queue.

• Agents work from database records: Datagrid imports operational records directly, so agents analyze transactional context from the synced copy.

• Controlled import scope: Project teams select the tables, views, and stored procedures agents need while limiting exposure across the instance.

• Production-safe reporting patterns: Agents execute checks and reports in Datagrid using the synced copy, including cross-system comparisons.

• Cross-system answers from one place: Agents join Cloud SQL records with data from 100+ connected platforms to detect mismatches and answer questions no single system can.

• Natural language access to relational data: Project teams ask questions in plain language, and agents interpret the synced tables to deliver answers in plain language.

• Managed database, managed extraction: Cloud SQL handles database operations such as backups and failover; Datagrid handles retrieval and downstream transformation.

What you can build with Google Cloud SQL - PostgreSQL Datagrid integration

The examples below show how synced Cloud SQL records become agent-executed workflows across reporting, validation, querying, and extraction.

• Recurring operational reports from transaction tables: Import order and invoice tables, or related project records. An agent assembles weekly status reports for the teams that need them and standardizes the reporting package.

• Cross-system record validation: An agent compares customer or vendor records in Cloud SQL against your CRM or warehouse data, then flags mismatched fields for review.

• Natural language querying for non-technical teams: Operators ask questions like which accounts had no activity last quarter. The agent interprets the synced tables and returns the answer, with no SQL required from the reader.

• AI-enriched extraction pipelines: An agent reads source tables and produces structured outputs for downstream workflows after applying cleaning and transformation logic, following the ETL replication model documented in Google Cloud SQL.

Resources and documentation

• Cloud SQL features page: Describes Cloud SQL for PostgreSQL capabilities.

• Authorized networks guide: Explains how to set up public IP access.

• Admin API: Documents programmatic instance administration.

• Import-export guide: Explains bulk data movement outside the integration.

Frequently asked questions

What authentication method does the Datagrid integration use?

Use a dedicated PostgreSQL database user with read access to the schemas that Datagrid should sync. Do not reuse an application account.

Does the integration write data back to my Cloud SQL instance?

No. Data flows in one direction, from Cloud SQL for PostgreSQL to Datagrid only.

How do I let Datagrid connect to my Cloud SQL instance?

For direct connections over public IP, follow the authorized-network setup, then test the connection during Start First Import.

Which PostgreSQL versions does Cloud SQL support?

Datagrid works with standard PostgreSQL credentials across Cloud SQL for PostgreSQL versions that Google currently lists as supported. Check the Cloud SQL FAQ before connecting to a new major version or planning a migration.

Can I use IAM database authentication instead of a password?

The Datagrid integration currently uses username and password credentials. For direct application connections where possible, Google recommends IAM database authentication in its connection options documentation. IAM tokens expire after one hour and require SSL, which suits Google's own tools and Auth Proxy rather than external credential-based integrations.

Similar integrations

• Azure PostgreSQL Database: The cross-cloud managed PostgreSQL counterpart for planning migrations and integration workflows between Azure and Google Cloud SQL PostgreSQL.

• BigQuery: An analytics destination for Cloud SQL PostgreSQL using Datastream or federated queries for ELT and near-real-time reporting.

• Google Cloud Storage: Supports SQL dump and CSV export-import workflows when moving data between Cloud SQL PostgreSQL and external systems.

• Amazon Redshift: A common cloud data warehouse destination for CDC and ELT pipelines ingesting Cloud SQL PostgreSQL data for analytics.

• Snowflake: A cloud data warehouse alternative used with CDC or ETL to centralize Cloud SQL PostgreSQL operational data for analytics and BI.

• PostgreSQL: The generic/self-hosted PostgreSQL integration, useful for migrating or syncing schemas and data alongside Cloud SQL-hosted databases.

Browse by category

• Data