Skip to main content

Managed PostgreSQL Preview

Concepts

Discover the fundamentals and essential principles to master our infrastructure.

Explore concepts →

Getting Started Guide

Get started quickly by following clear and simple instructions.

Launch the Quickstart →

Overview​

This product is in a preliminary version, and its documentation may contain errors or inaccuracies.

Managed PostgreSQL (on Kubernetes) by Cloud Temple is a managed PostgreSQL database engine solution hosted on Kubernetes. It complements the managed database engine offerings on virtual machines (referred to here as Managed PostgreSQL (on IaaS))

This product is suitable for customers who have Kubernetes workloads requiring PostgreSQL databases, or for customers who wish to consolidate multiple PostgreSQL database engines on a single Kubernetes cluster (consolidation). It is particularly well-suited for small and medium-sized databases that do not require tuning or specific features. For large-scale databases or those requiring specific tuning, it is recommended to choose the Managed PostgreSQL (on IaaS) product, which allows for more customization by our DBA expert teams.

Key Benefits​

  • Sovereignty and Reversibility : The solution relies exclusively on open source standards to avoid technological dependency and ensure the portability of your applications.
  • Simplicity and Delegation : The solution enables delegating to Cloud-Temple the management of database engines, specifically: updates and backups.

Deployment Models​

We offer two deployment models to meet your needs: StandAlone or Distributed.

StandAlone​

The StandAlone model deploys a single instance of the PostgreSQL engine in a multi-AZ infrastructure.

The storage used by this instance is replicated across 3 AZ, and allows automatic restart of the PostgreSQL instance on another AZ in case of failure.

  • Use case: This deployment model is perfectly suited for simple applications, such as CMSs, which use only a single endpoint to connect to databases.
  • Key points:
    • 1 database engine instance
    • storage distributed across 3 AZ for automatic failover in case of failure
    • physical and logical backups
    • 99.9% SLA (excluding maintenance windows)

Distributed​

The Distributed model deploys a cluster of 3 PostgreSQL database engine instances, with Patroni in "single primary" mode and PgBouncer:

  • a PgBouncer endpoint enables routing to the different instances based on the query type (read or write).
  • the read-write (RW) instance is accessible via a specific endpoint.
  • the read-only (RO) instances are accessible via another specific endpoint.

Thus, applications can choose to use RW or RO connections, or let PgBouncer automatically route to the most suitable endpoints.

  • Use case: This deployment model is ideally suited for applications with distributed access, such as data or business intelligence applications, which benefit from read-only access without impacting data ingestion.
  • Key points:
    • 3 database engine instances with Patroni in "single primary" mode
    • PgBouncer proxy for efficient query routing.
    • storage distributed across 3 AZs for automatic failover in case of failure
    • PiTR and Logical backups
    • SLA 99.9% (excluding maintenance windows)

Common Features​

Versions​

PostgreSQL engines can be selected from all supported versions (currently 14 to 18) See the "versioning policy" on the official PostgreSQL website

Backup​

2 types of backups are implemented:

  • Point in Time Restoration (PiTR): daily physical backup of the entire engine and its transaction logs (WAL), allowing the entire server to be restored to a specific past date without transaction loss.
  • Logical backup (pg_dump): logical exports of databases, for individual restore/export per database.

All backups use Cloud-Temple S3 storage (SNC-qualified) with at-rest encryption.