simple-gcp-connector 0.1.3

A simple connector for GCP databases

Low Overhead GCP Database Connector.

Low-overhead connections to Google Cloud SQL databases, optionally supporting IAM Authentication.

Prerequisites

• Google Cloud Project: A GCP project with Cloud SQL enabled.
• Cloud SQL Instance: A PostgreSQL instance configured for IAM authentication.

IAM Authentication (Optional)

• IAM User: A database user mapped to a Google Cloud IAM service account or user.
• Authentication:
  • Locally: gcloud auth application-default login
  • Production: Service Account attached to the environment (VM, GKE, Cloud Run).

Poastgres

1. Psycopg3 Connection Pool

Example using psycopg (version 3) to connect to Cloud SQL. It can automatically resolve the instance IP and inject the IAM token.

Dependencies:

• psycopg 3.3.0 or higher
• google-auth

import psycopg
from psycopg_pool import ConnectionPool
from simple_gcp_connector.psycopg import GoogleCloudConnInfoProvider

# Use the instance connection name (project:region:instance).
INSTANCE_CONNECTION_NAME = "my-project:us-central1:my-instance"
DB_USER = "[e-mail]" # note there is no .gserviceaccount.com
DB_NAME = "your-database-name" # likely postgres
DATABASE_URL = f"postgresql://{DB_USER}@IGNORED-HOST/{DB_NAME}?sslmode=require"

def create_pool():
    get_conninfo = GoogleCloudConnInfoProvider(
        DATABASE_URL,
        instance_connection_name=INSTANCE_CONNECTION_NAME
    )
    return ConnectionPool(
        conninfo=get_conninfo,
        min_size=1,
        max_size=5
    )

# Usage
with create_pool() as pool:
    with pool.connection() as conn:
        print("Connected!")
        # ... execute queries ...

2. SQLAlchemy

Use sqlalchemy to connect to Cloud SQL. If you are using the instance connection name, it will automatically resolve the instance IP and inject the IAM token unless enable_iam_auth is set to false (it's True by default).

Dependencies:

• sqlalchemy
• psycopg (or pg8000)
• google-auth

import sqlalchemy
from sqlalchemy import create_engine
from simple_gcp_connector.sqlalchemy import register_connector

# Configuration
INSTANCE_CONNECTION_NAME = "my-project:us-central1:my-instance"
DB_USER = "[e-mail]"
DB_NAME = "your-database-name"

engine = create_engine(
    f"postgresql+psycopg://{DB_USER}@/{DB_NAME}",
    connect_args={"sslmode": "require"}
)

register_connector(engine, instance_connection_name=INSTANCE_CONNECTION_NAME)

# Usage
with engine.connect() as connection:
    result = connection.execute(sqlalchemy.text("SELECT 1"))
    print(result.fetchone())

Note on Cloud SQL Proxy

If you are using the Cloud SQL Auth Proxy (sidecar or local process), you don't need instance_connection_name. You can simply point to 127.0.0.1.

# Psycopg w/ Proxy
get_conninfo = GoogleCloudConnInfoProvider("postgresql://user@127.0.0.1:5432/db")

# SQLAlchemy w/ Proxy
engine = create_engine("postgresql+psycopg://user@127.0.0.1:5432/db")
register_connector(engine)

Setup Notes

• Ensure the Service Account (or user) has the Cloud SQL Client role and is added as an IAM user in the Cloud SQL instance.
• If youa enable_iam_authentication flag must be set to on for your Cloud SQL instance.