ClickHouse Cloud: Enterprise Postgres Arrives

Unified Data Stack Evolves

ClickHouse's foray into managed Postgres is a bold move, aiming to bridge the OLTP/OLAP divide with a unified data stack. The article highlights key enterprise-grade features like cross-AZ HA with quorum-based replication, failover-safe CDC to ClickHouse, robust backup/PITR, and stringent security through dedicated VPCs. The emphasis on separating HA standbys from read replicas is a technically sound decision to prioritize durability and failover readiness, albeit a trade-off for immediate read scaling on HA instances. The integration with ClickHouse for real-time analytics via CDC is a strong differentiator. However, the initial focus on NVMe-backed storage and AWS suggests a potential limitation for users on other cloud providers or those with specific storage cost sensitivities, though GCP support is slated for the future. The lack of exposed read replicas on HA standbys means users will need to provision separate read replicas for scaling, adding a layer of management complexity and cost, which could be a concern for cost-conscious organizations.

The technical depth around replication, failover mechanisms, and security isolation is commendable. The use of WAL-G for backups and PITR is a well-established, robust choice. The failover-safe CDC slots are particularly innovative, addressing a common pain point in managed database services during HA events. The dedicated VPC per cluster is a significant security and isolation advantage over more shared multi-tenant architectures. The roadmap, including Query Performance Insights and broader cloud provider support, indicates a commitment to addressing these initial limitations. The target audience is clearly enterprises seeking a consolidated data platform that can handle both transactional and analytical workloads natively, reducing architectural complexity and data movement overhead. Developers and DevOps teams will find the focus on IaC support and intuitive SQL console appealing for operational efficiency.

Key Points

• ClickHouse is launching a managed Postgres service integrated within ClickHouse Cloud, aiming for an always-available, reliable, and observable OLTP offering.
• Key features include cross-AZ High Availability (HA) with quorum-based synchronous replication (1 or 2 standbys), prioritizing durability over read scaling on HA standbys.
• Built-in failover-safe CDC leverages ClickPipes/PeerDB and replication slots that persist across HA events, preventing resyncs when replicating data to ClickHouse.
• Robust backup and Point-In-Time Recovery (PITR) capabilities are provided using WAL-G for continuous WAL archiving to object storage.
• Security is enhanced with dedicated VPCs per cluster for network isolation, IP allow-listing, TLS encryption for data in transit, and AES-256 encryption for data at rest.
• Monitoring includes core system and database metrics, with plans for Query Performance Insights and more comprehensive logging.
• Future roadmap includes GCP support, CDC v2 for sub-second replication, storage autoscaling, IaC support (OpenAPI, Terraform), and improved operational visibility.

---

📖 Source: Building an enterprise Postgres service in ClickHouse Cloud