HOWTO: Provision a Cloud SQL Standby for PostgreSQL DR¶

Purpose: Provision a Cloud SQL PostgreSQL instance as a managed DR standby, configure the replication feed, and validate it against the DR decision service.

Difficulty: Advanced

Track: Disaster Recovery Automation

Overview¶

Cloud SQL in HybridOps is not a general database service: it is a specific DR lane with defined RTO/RPO targets, a wired health signal, and a tested failover path. This HOWTO covers provisioning the standby instance and confirming it is ready to receive a failover decision.

1. Architecture Position¶

On-prem primary → Cloud SQL logical replica topology.
Cloud SQL DR lane vs self-managed pgBackRest lane.
RTO/RPO targets for the Cloud SQL path.

2. Module Overview¶

Module ID: gcp/cloudsql-dr-standby.
GCP project, region, and instance tier inputs.
Terraform provider configuration.

3. Provisioning the Cloud SQL Instance¶

Module input map: instance spec, database version, network configuration.
VPC peering or Private Service Connect for on-prem connectivity.
Module execution and run record review.

4. Configuring Logical Replication¶

Publication on the on-prem primary.
Subscription on Cloud SQL.
Replication lag monitoring.

5. Health Signal Wiring¶

Cloud SQL health check endpoint registration in the DR decision service.
Replication lag signal threshold configuration.
Test signal emission and decision service response.

6. Pre-Drill Validation¶

Replication lag below RPO threshold.
Application credentials provisioned on Cloud SQL.
DNS failover entry pre-configured.

References¶

License: MIT-0 for code, CC-BY-4.0 for documentation unless otherwise stated.