The Hybrid Data Pipeline Helm chart provides support for the PostgreSQL replication architecture which enables high availability and data resilience through a primary-replica configuration. This feature allows you to deploy PostgreSQL with multiple read replicas for improved fault tolerance and data protection.

PostgreSQL replication provides a primary instance that handles all write operations and multiple read replicas that maintain copies of the data for recovery. Synchronous replication ensures data is safely written to replicas before confirming success to the application.

By default, PostgreSQL is deployed in standalone mode with a single instance. You can enable replication mode by configuring the architecture setting and replica count in the values.yaml file.

PostgreSQL replication can be configured in the values.yaml file with the following settings:

  • Enable replication mode instead of standalone:
    
  postgresql:
    architecture: replication
  • Configure synchronous commit and replica behavior:
    
  postgresql:
    replication:
      synchronousCommit: "remote_apply"
      numSynchronousReplicas: 1
      applicationName: hdp
  • Set replica count and resource allocation
    
  postgresql:
    readReplicas:
      name: read
      replicaCount: 2
      resources:
        requests:
          cpu: 2
          memory: 4Gi
  • Connect primary to external PostgreSQL clusters:
    
  postgresql:
    primary:
      standby:
        enabled: true
        primaryHost: "external-postgres-primary.example.com"
        primaryPort: "5432"

For descriptions of supported parameters, see Helm chart parameters or refer to the values.yaml file in the Hybrid Data Pipeline Helm chart GitHub repository.