As we build [Postgres managed by ClickHouse](https://clickhouse.com/cloud/postgres), an NVMe-backed Postgres service native integrated with ClickHouse, our top priority (P0) is delivering an always-available, reliable, and operationally observable managed Postgres offering. We believe this is table stakes for OLTP workloads!

With that vision in mind, we’ve been investing heavily in state-of-the-art features to ensure customers get an enterprise-grade Postgres experience. In this blog, we highlight several key features we’ve shipped over the past few months and share a preview of what’s ahead on our roadmap to further strengthen platform maturity and enterprise readiness. Let’s get started!

## Cross AZ HA

Postgres managed by ClickHouse supports up to two synchronous standbys spread across availability zones, protecting you from node, rack, or full AZ failures. With 2 standbys configured, we use [quorum-based streaming replication](https://www.postgresql.org/docs/current/runtime-config-replication.html#GUC-SYNCHRONOUS-STANDBY-NAMES): a write is acknowledged when at least one standby confirms it. This provides strong durability guarantees, without turning every commit into a latency tax. In the event of a failure, failover is automatic, promoting a standby with minimal disruption.

![quorum-based streaming replication](/uploads/552707022_938b77d4_1c8d_4fae_ab97_c0ed6edacee2_92b4642245.jpg)

For cost-sensitive workloads that require high availability, but aren’t tier 0, you have the flexibility to choose 1 cross-AZ standby. It uses synchronous replication to provide strong availability, while offering lesser fault tolerance than a two-standby quorum configuration.

> An important architectural nuance: our HA standbys are not exposed for reads, ensuring that we prioritize failover readiness and data durability over opportunistic read scaling.

HA standbys are reserved strictly for failover and are not exposed for read traffic. Read traffic on standbys can compete with WAL replay, increasing replication lag, and can delay failover readiness. Long-running queries on replicas can also interfere with VACUUM and bloat control on primary. We avoid those trade-offs by keeping standbys focused solely for HA. If you need read scaling, **we provide separate read replicas designed specifically for that purpose**.

## Reliable CDC to ClickHouse (failover-safe slots)

The service comes with native ClickHouse integration with built-in Change Data Capture (CDC), enabling continuous replication of transactional data into ClickHouse for real-time analytics. The integration is powered by ClickPipes/PeerDB, a battle tested replication engine supporting 100s of Postgres customers.

> A unique reliability feature of the service is built-in failover replication slots, preventing resyncs on primary failover.

In most managed Postgres services, logical replication slots are tied to the primary instance. During high-availability failovers, maintenance events, or scaling operations, these slots can be lost or require manual recreation, interrupting CDC pipelines and potentially forcing full re-syncs.

Postgres by ClickHouse includes built-in infrastructure for failover replication slots when syncing data to ClickHouse. These slots are preserved across HA failovers and scaling operations. As a result, CDC pipelines continue running without manual intervention or slot re-creation when the primary changes, reducing the risk of costly resyncs on large databases.

## Backups, PITR and Forks

High availability protects you from infrastructure failures. Backups protect you from everything else.

Every Postgres service includes **automatic backups** with support for **point-in-time recovery (PITR)** and **forks**. We use **WAL-G**, a widely adopted open-source tool, to take full base backups and continuously archive WAL to object storage (S3-compatible). WAL-G is well tuned to perform full backups and restores, as well as WAL archival and retrieval, in parallel to meet the high throughput demands of large-scale workloads. We also use the [`wal-g` daemon](https://github.com/wal-g/wal-g/blob/master/docs/PostgreSQL.md#daemon), as it runs as a persistent process, eliminating per-WAL process startup overhead and enabling efficient, low-latency, and reliable WAL shipping under sustained write volumes.

backups_e.mp4


Continuous WAL archiving allows you to restore your database to an exact timestamp within the retention window (default 7 days), not just to the last snapshot. Forks are built on the same mechanism. You can create a new Postgres service from any recovery point, making it easy to test migrations, debug issues, or run backfills without impacting production.

Together with HA, backups provide a second layer of protection, covering not just infrastructure failures, but human error and application bugs as well.

## Zero compromise security

Postgres managed by ClickHouse is designed with practical, production-grade security controls across network access, encryption, and isolation.

> A key architectural differentiator is that **every Postgres cluster runs inside its own dedicated, isolated VPC.** 

This provides hard network-level isolation between customer environments, with no shared runtime and no cross-database interference. The result is strong blast-radius containment and reduced multi-tenant risk. While many managed database services rely on container orchestration platforms with shared networking layers, we provision dedicated environments per cluster to deliver a higher level of separation by default.

![dedicated, isolated VPCs](/uploads/552707023_277c2c51_35a9_4343_9a77_77ee0d1edca1_14eec91d6a.jpg)

At the network layer, access can be restricted using [IP allow-listing](https://clickhouse.com/docs/cloud/managed-postgres/security#ip-whitelisting), ensuring that only approved application servers, VPN endpoints, bastion hosts, or CI/CD systems can connect to the database.

All data is [encrypted](https://clickhouse.com/docs/cloud/managed-postgres/security#encryption) both in transit and at rest. Client connections are secured with TLS (SSL) to protect against interception and man-in-the-middle attacks. Encrypted connections can be strictly enforced, and for environments requiring stronger authentication guarantees, certificate-based client verification can be used. Data at rest, including NVMe storage and object storage used for backups and WAL archives,  is encrypted using AES-256, the industry-standard encryption algorithm widely adopted across cloud providers. Encryption keys are securely managed by the underlying cloud infrastructure.

For environments requiring stronger network isolation, [Private Link support](https://clickhouse.com/docs/cloud/managed-postgres/security#private-link) enables private connectivity between your VPC and the Postgres service, ensuring that database traffic does not traverse the public internet.

## Monitoring

The Postgres service comes with built-in, out-of-the-box metrics that provide visibility into both system health and database performance. Today, it includes 10+ core metrics, allowing you to monitor CPU usage, memory consumption, disk I/O, storage growth, and network traffic, alongside Postgres-specific signals such as connections, transaction throughput, cache hit ratios, row activity, and deadlocks.

monitoring_e.mp4

We’re actively expanding operational visibility, including Query Performance Insights and more comprehensive logging.

## Looking Ahead

As we execute on **our vision of delivering a unified data stack for OLTP and OLAP by combining Postgres and ClickHouse**, providing an enterprise-grade experience is a given!

In the coming quarters, we’re expanding the platform’s enterprise readiness across reliability, performance, and operational depth. A few features include support for additional cloud providers (GCP), CDC v2 for close to sub-second replication into ClickHouse, and Query Performance Insights (QPI) for deeper visibility into query behavior and database internals.

We’re also investing in core platform capabilities such as storage autoscaling, OpenAPI and Terraform support for IaC, one-click fully managed Postgres-to-Postgres migrations, improved operational visibility (e.g., scaling progress), a built-in intuitive SQL console and more.

Stay tuned.

ClickHouse + Postgres has become the unified data stack for applications that scale. With Managed Postgres now available in ClickHouse Cloud, this stack is a day-1 decision.

Try Postgres managed by ClickHouse

We’ve been investing heavily in state-of-the-art features to ensure customers get an enterprise-grade Postgres experience.

large_552217859-2606de73-2445-4ff5-886c-27321950cc73.jpg

medium_552217859-2606de73-2445-4ff5-886c-27321950cc73.jpg

small_552217859-2606de73-2445-4ff5-886c-27321950cc73.jpg

thumbnail_552217859-2606de73-2445-4ff5-886c-27321950cc73.jpg

552217859-2606de73-2445-4ff5-886c-27321950cc73.jpg

Building an enterprise Postgres service in ClickHouse Cloud

Today, we're announcing a fast, scalable and an enterprise-grade managed Postgres service natively integrated with ClickHouse, built for real-time and AI-driven applications. We are offering a **Unified Data Stack** that brings transactional and analytical workloads together without traditional complexity. By combining Postgres for transactions with ClickHouse for analytics, teams get best-in-class performance and scalability on an open-source foundation while significantly reducing operational overhead.

Developers get a high-performance, scalable Postgres service backed by local NVMe storage, delivering up to 10X faster performance for fast-growing workloads that are mostly disk-bound. In just a few clicks, using native CDC capabilities, they can sync transactional data in Postgres to ClickHouse to make it analytics-ready, unlocking up to [100X faster analytics](https://benchmark.clickhouse.com/#system=+lik%7CgQ&type=-&machine=-ca2l%7C6t%7Cg4e%7C6ax%7Cae-l%7C6ale%7Cg-l%7C3al&cluster_size=-&opensource=-&hardware=-&tuned=+n&metric=combined&queries=-). With a unified query layer powered by the [pg_clickhouse](https://github.com/ClickHouse/pg_clickhouse) Postgres extension, they can build applications that seamlessly span transactions and analytics without managing separate systems.

To try ClickHouse's native Postgres service, sign up for Private Preview using this link.

Get started with our native Postgres Service

This demo showcases how the various pieces come together to form a Unified Data Stack powered by Postgres and ClickHouse.

<iframe width="768" height="432" src="https://www.youtube.com/embed/rpBA13nQxAk?si=Xu1e9FERMcnXgKWh" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>

To achieve this, we're partnering with [Ubicloud](https://www.ubicloud.com/), an open-source cloud company that offers one of the fastest Postgres services and whose [founding team](https://www.ubicloud.com/about-ubicloud) has a track record of delivering world-class Postgres at Citus Data, Heroku, and Microsoft.


> Noah Pryor, the CTO of [Beehiiv](https://www.beehiiv.com/), a large ClickHouse customer, shared excitement for this launch, "We’re excited to see ClickHouse entering the Postgres ecosystem. At Beehiiv, we rely heavily on both Postgres and ClickHouse to power our mission-critical, customer-facing applications. We’ve invested significant effort in integrating these technologies, so a tighter, more native integration between them would materially simplify our architecture."

## Why Postgres managed by ClickHouse

### Postgres, built for performance and scale

Across thousands of fast growing AI-driven companies using Postgres, performance and scalability consistently emerge as challenges. Common issues include slower ingestion (mainly updates), slower vacuums, increased tail latency for reads and writes, long-running, slow transactions causing lock contention and sudden WAL spikes, among others. In most cases, these issues stem from limited disk IOPS and suboptimal disk latency. With unlimited IOPS, Postgres can operate more efficiently than it does today on most other managed Postgres services.

Postgres managed by ClickHouse is an enterprise-grade managed Postgres service backed by NVMe storage. The NVMe storage is physically colocated with compute, delivering orders of magnitude lower disk latency and higher IOPS than alternatives such as EBS-backed volumes, which require a network round trip for disk access. As a result, disk access latency can be on the order of microseconds rather than milliseconds. For fast-growing Postgres workloads that are primarily throttled by disk IOPS and latency, we expect up to 10X speedups at the same cost and hardware profile compared to other Postgres providers.

> *"We're excited that ClickHouse is offering a native Postgres service to complement its core OLAP database. NVMe-backed Postgres performance, combined with seamless integration with CDC and pg_clickhouse for fast analytics, makes it very compelling for us to build our fast-growing applications."* - Aditya Maru, Co-founder, [Blacksmith.sh](http://blacksmith.sh)

The image below shows the reduction in P99 latency for disk-bound workloads when moving from traditional Postgres to NVMe-backed Postgres.

![postgres-launch-1.png](/uploads/postgres_launch_1_5e02d22e1a.png)

#### Ubicloud's NVMe-backed Postgres stood out

To select the best architecture, we evaluated many Postgres providers as potential partners. Ubicloud stood out by offering a reliable Postgres service backed by NVMe storage that addressed the same customer challenges we observed and outperformed other providers by orders of magnitude in real-world benchmarks. We plan to share a detailed blog post on our evaluation process in the future.

We evaluated alternative architectures, including separating storage and compute, and found that object-storage-backed Postgres is not well suited for OLTP workloads, where consistent  and predictable latency and throughput are essential. Object storage works well for analytics, but its higher latency and coarse-grained I/O make it challenging to support transactional patterns. We’ll dive deeper into this trade-off in a follow-up blog post.

### Deeply integrated with ClickHouse

Developers can natively integrate their Postgres with ClickHouse to enable 100x faster analytics. It comes with two critical features enabling this integration. These features are central to powering the Unified Data Stack, combining Postgres for transactions and ClickHouse for analytics with no overhead.

The diagram below depicts the Unified Data Stack, highlighting Postgres for OLTP, ClickHouse for OLAP, continuous replication between the two, and a unified query layer via the `pg_clickhouse` extension. 

![postgres-launch-2.png](/uploads/postgres_launch_2_5b8721a622.png)

### Postgres to ClickHouse replication

Seamlessly replicate Postgres data to ClickHouse in real time to make it analytics-ready. We support both an initial load to migrate existing data to ClickHouse and CDC for incremental syncs. This feature is powered by the Postgres CDC connector in [ClickPipes](https://clickhouse.com/docs/integrations/clickpipes/postgres), which is battle-tested by 100s of enterprise customers moving 100s of TB per month.

We plan to offer additional features for more native CDC, including sub-second replication latency, replication of ongoing transactions using logical replication v2 for reliable slot flushing, etc. all of which will be exclusive to our native Postgres service.

### pg_clickhouse: unified query layer for OLTP and OLAP

Every Postgres service comes with the [pg_clickhouse](https://clickhouse.com/blog/introducing-pg_clickhouse) Postgres extension, which enables users to query ClickHouse directly from Postgres. It allows Postgres to act as a unified query layer, powering not just transactions but also analytics using ClickHouse internally, making it easy for developers to build applications.

`pg_clickhouse` provides comprehensive query pushdown to ClickHouse for efficient query execution, including support for FILTERs, JOINs, SEMI-JOINs, aggregations, functions and more. Currently, 14 of 22 TPC-H queries are fully pushed down, delivering over 60x performance improvements compared to standard Postgres.

We will continue expanding pushdown support to more complex queries, including CTEs, window functions, and beyond, all with the goal of enabling users to build fast analytics from the Postgres layer itself while still leveraging the power of ClickHouse.

Users can seamlessly create foreign tables in Postgres that point to ClickHouse tables under the hood and start powering analytics using these foreign tables. As we evolve `pg_clickhouse`, we'll work toward making this experience as UX-friendly as possible, including auto-creating foreign tables for replicated data, auto-routing transactions and analytics to the appropriate engines (Postgres or ClickHouse), and more.

### Enterprise-grade Postgres

#### Reliability without compromise

Our native Postgres service provides enterprise-grade reliability from day one. We offer strong high availability (HA), allowing a primary with up to two standby instances across different availability zones (AZs) using quorum-based replication. These standby instances are dedicated to HA and auto-failover; they are not exposed as read replicas, as that could impact HA guarantees. Read scaling is handled via separate read replicas (details [here](https://clickhouse.com/docs/cloud/managed-postgres/read-replicas)).

For disaster recovery, every service comes with automatic backups that support forks and point-in-time recovery. The backups run on [WAL-G](https://github.com/wal-g/wal-g), a well-known open-source Postgres tool that handles full backups and continuous WAL archiving to object storage like S3. WAL-G was [originally created](https://www.citusdata.com/blog/2017/08/18/introducing-wal-g-faster-restores-for-postgres/) by Daniel Farina, co-founder of Ubicloud, during his time at Citus Data, another testament to the quality of the team we're partnering with.

#### Built on ClickHouse Cloud trust

It is built to meet the same high standards for security, privacy, and compliance as ClickHouse Cloud. As part of a unique partnership, Ubicloud runs entirely within AWS accounts owned by ClickHouse Cloud and is subject to the same security, privacy, and governance controls applied across the platform. It already includes core capabilities such as SAML/SSO, IP allow-listing, data encryption at rest and in transit, as well as PrivateLink (available via a support ticket), with plans to further align and unify with the advanced features offered by ClickHouse Cloud over time.

### Open source first

The managed Postgres service is built on an open source first philosophy. [Postgres](https://github.com/postgres/postgres) and [ClickHouse](https://github.com/ClickHouse/ClickHouse) are both open-source databases with large, thriving communities. Every component that integrates Postgres and ClickHouse is open source, from the [pg_clickhouse](https://github.com/ClickHouse/pg_clickhouse) extension that enables a unified query layer to [PeerDB](https://github.com/PeerDB-io/peerdb), which powers real-time CDC and replication.

[Ubicloud](https://github.com/ubicloud/ubicloud), our strategic partner on the managed Postgres service, is also an open source company, deeply aligned with this philosophy This was a major factor in choosing the right partner. We are jointly contributing to the open-source ecosystem to deliver a world-class Postgres to the community, and this collaboration is already underway. [Here](https://github.com/ubicloud/ubicloud/pulls?q=is%3Apr+author%3Aserprex) are a few key contributions from the ClickHouse team to the Ubicloud open-source repository.

This open source foundation ensures there is no vendor lock-in, giving developers full control, transparency, and long-term flexibility over their data stack.

## Try Postgres managed by ClickHouse



Our team will reach out within a day and work closely with you to provide early access. **The private preview is fully free and entails no cost.**

It is currently available on AWS in 10 public regions, with over 50 NVMe-backed configurations ranging from 2 vCPUs / 8 GB RAM / 118 GB disk to 96 vCPUs / 768 GB RAM / 60 TB storage. We plan to expand to additional regions and other cloud service providers, including GCP and Azure, in the near future. The offering includes all standard managed service features, including high availability, forks, read replicas, PgBouncer, upgrades, and more, along with deep integration with ClickHouse.

ClickHouse launch a fast, scalable and enterprise-grade managed Postgres service natively integrated with ClickHouse, built for real-time and AI-driven applications.

large_postgres-launch-banner.png

medium_postgres-launch-banner.png

small_postgres-launch-banner.png

thumbnail_postgres-launch-banner.png

postgres-launch-banner.png

Fast, scalable, enterprise-grade Postgres natively integrated with ClickHouse

Seamlessly connect your external data sources to ClickHouse Cloud - try out ClickPipes today!

Get started today

A look back at the evolution of the Postgres CDC connector in ClickHouse Cloud over its first year, following the 2024 PeerDB acquisition. 

large_Postgres CDC ClickHouse Review.jpg

medium_Postgres CDC ClickHouse Review.jpg

small_Postgres CDC ClickHouse Review.jpg

thumbnail_Postgres CDC ClickHouse Review.jpg

Postgres CDC ClickHouse Review.jpg

Postgres CDC in ClickHouse, A year in review

本日、ClickPipes向け [Postgres CDCコネクタ](https://clickhouse.com/cloud/clickpipes/postgres-cdc-connector)の一般提供開始を発表できることを嬉しく思います。これにより、お客様はPostgresデータベースをClickHouse Cloudへ数クリックで簡単にレプリケートできるようになります。

この発表は、大きなマイルストーンです。[PeerDB](https://www.peerdb.io/) ― Postgres CDCに特化した企業であり、昨年[ClickHouseが統合](https://clickhouse.com/blog/clickhouse-welcomes-peerdb-adding-the-fastest-postgres-cdc-to-the-fastest-olap-database)したパートナー ― は、ClickPipes内のPostgres CDCコネクタとしてClickHouse Cloudに完全統合されました。このコネクタは、エンタープライズレベルのPostgresユースケースにも対応可能です。

> [ClickHouse Cloudにサインアップ](https://console.clickhouse.cloud/signup)して、[ClickPipes向けPostgres CDCコネクタ](https://clickhouse.com/docs/integrations/clickpipes/postgres)をお試しください。 

## Postgres + ClickHouse = 「デフォルトのデータスタック」

近年、[GitLab](https://about.gitlab.com/blog/2022/04/29/two-sizes-fit-most-postgresql-and-clickhouse/)、[CloudFlare](https://blog.cloudflare.com/http-analytics-for-6m-requests-per-second-using-clickhouse...)、[Instacart](https://tech.instacart.com/real-time-fraud-detection-with-yoda-and-clickhouse-bd08e9dbe3f4) などの企業をはじめ、多くのビジネスで共通のパターンが見られるようになっています。それは、PostgresとClickHouseを組み合わせて、あらゆるデータ課題を解決するというものです。

* **Postgres** はトランザクション型Webアプリケーションのためのシステム・オブ・レコードとして機能します。
* **ClickHouse** はリアルタイム分析とレポート用途のためのシステム・オブ・アナリシスとして機能します。

このパターンはAI時代に入りさらに加速しており、[LangChain](https://clickhouse.com/blog/langchain-why-we-choose-clickhouse-to-power-langchain)、[LangFuse](https://langfuse.com/blog/2024-12-langfuse-v3-infrastructure-evolution)、[Vapi](https://neon.tech/blog/vapi-voice-agents-neon) などの企業も同様のアーキテクチャを採用しています。私たちは、Postgres + ClickHouseがモダンビジネスにおける「標準的なデータスタック」になりつつあると考えています。その統合を“魔法のようにシームレス”にすることが私たちの目標です。そして今回のPostgres CDCコネクタは、そのための**最初の大きなステップ**です。PostgresのデータをClickHouseへ簡単に取り込むことで、リアルタイム分析をスムーズに実現できるようになります。

以下は、PostgresとClickHouseをPostgres CDCで組み合わせた際のリファレンスアーキテクチャの図です。Postgresは低レイテンシなトランザクションを処理し、ClickHouseが高速な分析を支えます。

<img preview="/uploads/Postgres_Click_House_reference_architecture_61fe52239a.png"  src="/uploads/Postgres_Click_House_reference_architecture_61fe52239a.png" alt="catalogue_lakehouse.png" class="h-auto w-auto max-w-full"  style="width: 100%;">

## メトリクスと導入企業の事例

過去6か月間、このコネクタはユーザーフィードバックをもとに急速に進化を遂げながら、広範なベータフェーズを経てきました。現在では、**数百のミッションクリティカルなワークロード**を支え、**月間100TB以上のデータ**をClickHouseに移行しています。

代表的な導入企業には、米国の大手採用プラットフォーム [**Ashby**](https://www.ashbyhq.com/)、急成長中のサイバーセキュリティスタートアップ [**Seemplicity**](https://seemplicity.io/)、米国有数の自動車小売企業 [**AutoNation**](https://www.autonation.com/) などがあります。以下は、一部のお客様からの声です。

> 「ClickHouseはAshbyの顧客向け分析基盤として稼働しており、完全に動的なインサイトを超高速で提供しています。一方でPostgresはコアトランザクションを処理しています。ClickPipes経由のPostgres CDCにより、テラバイト規模のデータをシームレスにレプリケートでき、リアルタイム分析がさらに高速化しました。以前は数分かかっていたレポートが、今では1秒以内で完了します。エンタープライズ顧客がリアルタイムで意思決定する際に活用するダッシュボードは、まさに“真実の情報源”です。PostgresとClickHouseを組み合わせることで、私たちは信頼性の高い体験と、スケールに応じたデータ処理を提供できています。」
> — [Elenie Godzaridis](https://www.linkedin.com/in/elenie-godzaridis/), Director of Engineering, [Ashby](https://www.ashbyhq.com/)

> 「最初はDebeziumを使って社内でPostgres CDCを実装しようとしましたが、あまりに複雑すぎました。PostgresのビットをClickHouseに変換することを使命にしているエンジニアたちが作ったマネージド製品のほうが、自分たちの手で作るものよりも優れていると確信していました。」
> — [Tal Shargal](https://www.linkedin.com/in/tal-shargal-29388671/), Chief Architect, [Seemplicity](https://seemplicity.io/), [*導入事例はこちら*](https://clickhouse.com/blog/seemplicity-scaled-real-time-security-analytics-with-postgres-cdc-and-clickhouse)

![Chart.png](https://cms.clickhouse-dev.com:1337/uploads/Chart_7c2a541127.png)

**このグラフは、過去1年間におけるPostgres CDCからClickHouseへの使用量の成長を示しています。** [**リファレンス**](https://www.linkedin.com/feed/update/urn:li:activity:7310749593296588800/)

## 製品の特長

<iframe width="660" height="371" src="https://www.youtube.com/embed/s1hREjN02S0?si=J2ByKIkTTU9Mw94m" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>

Postgres CDCコネクタは、PostgresとClickHouseのために専用設計されたコネクタであり、豊富な機能を備えています。以下は特に注目すべき機能の一部です：

1. **初回ロードおよび再同期が10倍高速化** –  [並列スナップショット](https://blog.peerdb.io/parallelized-initial-load-for-cdc-based-streaming-from-postgres)により、大規模な単一テーブルを並列に読み込み可能。テラバイト規模のデータでも、従来の「数日」ではなく「数時間」で移行できます。

2. **最小10秒のレプリケーション遅延** –  レプリケーションスロットは[再接続なしで継続的に消費](https://clickhouse.com/blog/enhancing-postgres-to-clickhouse-replication-using-peerdb#efficiently-flush-the-replication-slot)され、ClickHouseの[ReplacingMergeTree](https://clickhouse.com/blog/postgres-to-clickhouse-data-modeling-tips-v2)がエンドツーエンドの遅延を最小化します。

3. **スキーマ変更の自動レプリケーション** –  [ADD COLUMNやDROP COLUMNといった操作](https://clickhouse.com/docs/integrations/clickpipes/postgres/schema-changes)にも対応し、手動による対応は不要です。

4. **テーブル・カラム単位の除外設定** –  PII（個人情報）保護やネットワークスループットの最適化のために、きめ細かな制御が可能です。

5. **セキュアな接続** –  [**AWS PrivateLink**](https://clickhouse.com/docs/integrations/clickpipes/aws-privatelink) や [**SSHトンネリング**](https://clickhouse.com/docs/integrations/clickpipes/postgres#optional-setting-up-ssh-tunneling) に対応しており、ソースのPostgresデータベースと安全かつプライベートに接続できます。

6. **Postgresのネイティブ機能に対応** –  [パーティションテーブル](https://blog.peerdb.io/real-time-change-data-capture-for-postgres-partitioned-tables)、TOASTカラム、[配列やJSONなどの高度なデータ型](https://clickhouse.com/docs/integrations/clickpipes/postgres/faq#how-are-postgres-data-types-mapped-to-clickhouse)のレプリケーションもサポートしています。

7. **Open API（ベータ）** –  [APIやCLIによるパイプの作成・管理](https://clickhouse.com/docs/integrations/clickpipes/postgres/faq#can-clickpipe-creation-be-automated-or-done-via-api-or-cli)が可能で、ClickHouse Cloud + ClickPipes環境を個別に構築・運用するISV（独立系ソフトウェアベンダー）にも最適です。[**Terraform**](https://registry.terraform.io/providers/ClickHouse/clickhouse/3.2.0-alpha1/docs/resources/clickpipe) **によるIaC対応も近日提供予定です。**

## 料金について

一般提供（GA）の開始に伴い、[ベータ期間中にもお知らせしていた](https://clickhouse.com/blog/postgres-cdc-connector-clickpipes-public-beta#pricing)とおり、Postgres CDCコネクタの料金体系を導入します。私たちの目標は、PostgresデータベースをClickHouseにシームレスかつ手頃な価格で接続できるようにするというビジョンを保ちながら、非常に競争力のある価格を実現することです。

このコネクタは、外部のETLツールや他のデータベースプラットフォームにある類似機能と比べて、**5倍以上のコスト効率**を誇ります。

なお、Postgres CDC ClickPipesを利用している**すべてのお客様（既存・新規）に対して、2025年9月1日より月次課金が開始**されます。必要に応じてコストを最適化するための**3か月間の猶予期間**が設けられていますが、多くのお客様にとって設定の変更は不要と想定しています。料金体系の詳細、具体的な例、よくある質問については[こちら](https://clickhouse.com/docs/cloud/manage/billing/overview#clickpipes-for-postgres-cdc)をご覧ください。

## Postgres CDC を使って ClickHouse にデータを取り込む方法

以下のリンクから、PostgresデータベースをClickHouse Cloudに接続して、超高速な分析を体験してみてください。

* [PostgresからClickHouseへのデータ取り込み（CDCを使用）](https://clickhouse.com/docs/en/integrations/clickpipes/postgres)

* [ClickPipes for Postgres よくある質問（FAQ）](https://clickhouse.com/docs/en/integrations/clickpipes/postgres/faq)

* [ClickHouse Cloudを無料で試す](https://clickhouse.com/docs/en/cloud/get-started/cloud-quick-start)



ClickPipes向けPostgres CDCコネクタの一般提供を開始しました。ClickHouse Cloudにネイティブに統合されており、Postgresデータベースのレプリケーションを数クリックで高速に実現できます。

large_Postgres CDC connector for ClickPipes is now Generally Available.png

medium_Postgres CDC connector for ClickPipes is now Generally Available.png

small_Postgres CDC connector for ClickPipes is now Generally Available.png

thumbnail_Postgres CDC connector for ClickPipes is now Generally Available.png

Postgres CDC connector for ClickPipes is now Generally Available.png

ClickPipes向けPostgres CDCコネクタが一般提供開始

Today we are thrilled to announce that the [Postgres CDC connector](https://clickhouse.com/cloud/clickpipes/postgres-cdc-connector) for ClickPipes is now generally available. With this customers can easily replicate their Postgres databases to ClickHouse Cloud within just a few clicks.

This announcement represents a major milestone. [PeerDB](https://www.peerdb.io/) — a leading Postgres CDC company we [joined forces](https://clickhouse.com/blog/clickhouse-welcomes-peerdb-adding-the-fastest-postgres-cdc-to-the-fastest-olap-database) with last year, is now fully integrated into ClickHouse Cloud as the Postgres CDC connector in ClickPipes. The connector is ready to support enterprise-grade Postgres use cases.  
  
>[Sign up for ClickHouse Cloud](https://console.clickhouse.cloud/signup) today to try out the [Postgres CDC connector for ClickPipes](https://clickhouse.com/docs/integrations/clickpipes/postgres)! 

<iframe width="660" height="371" src="https://www.youtube.com/embed/pwOEYtNtk9k?si=02uMFJwzoGLUODqo" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>

## Postgres + ClickHouse = “the default data stack”

Over the past few years, we are seeing a consistent pattern emerge across businesses incl. [GitLab](https://about.gitlab.com/blog/2022/04/29/two-sizes-fit-most-postgresql-and-clickhouse/), [CloudFlare](https://blog.cloudflare.com/http-analytics-for-6m-requests-per-second-using-clickhouse?glxid=25d1a1c5-1ce0-4589-8305-177654961925&pagePath=%2Fblog%2Fclickhouse-welcomes-peerdb-adding-the-fastest-postgres-cdc-to-the-fastest-olap-database&origPath=%2Fblog%2Fclickhouse-acquires-peerdb-to-boost-real-time-analytics-with-postgres-cdc-integration&utm_ga=GA1.1.1147851049.1729970044/) and [Instacart](https://tech.instacart.com/real-time-fraud-detection-with-yoda-and-clickhouse-bd08e9dbe3f4): Using Postgres and ClickHouse together to solve most of their data challenges.

* **Postgres** serves as the system of record powering transactional web apps.
* **ClickHouse** powers real-time analytical and reporting workloads, serving as the system of analysis.

This pattern is accelerating in the AI era, with companies like [LangChain](https://clickhouse.com/blog/langchain-why-we-choose-clickhouse-to-power-langchain), [LangFuse](https://langfuse.com/blog/2024-12-langfuse-v3-infrastructure-evolution) and [Vapi](https://neon.tech/blog/vapi-voice-agents-neon) adopting the same architecture. We believe Postgres + ClickHouse is becoming the default data stack for modern businesses. We're committed to making that integration magical. The Postgres CDC connector is **the first major step**—enabling seamless, real-time analytics by bringing your Postgres data into ClickHouse effortlessly. 

The image below shows the reference architecture for combining Postgres and ClickHouse using Postgres CDC, with Postgres handling low-latency transactions and ClickHouse powering blazing-fast analytics.

<img preview="/uploads/Postgres_Click_House_reference_architecture_61fe52239a.png"  src="/uploads/Postgres_Click_House_reference_architecture_61fe52239a.png" alt="catalogue_lakehouse.png" class="h-auto w-auto max-w-full"  style="width: 100%;">

## Some metrics and reference customers

Over the past six months, the connector has gone through an extensive beta phase, evolving rapidly based on user feedback. It now supports **hundreds of mission-critical workloads**, moving **100TB+ of data per month** into ClickHouse.

A few reference customers include. [**Ashby**](https://www.ashbyhq.com/)—a top recruitment company in the US, [**Seemplicity**](https://seemplicity.io/) – one of the fastest-growing cybersecurity startups, and [**AutoNation**](https://www.autonation.com/) – a well-renowned automotive retailer in the US. Here are a few testimonials from our reference customers.  

>“ClickHouse powers Ashby’s customer-facing analytics, delivering lightning-fast fully dynamic insights, while Postgres handles core transactions. With Postgres CDC via ClickPipes, we seamlessly replicated terabytes of data—speeding up our real-time analytics without disrupting operations. Reports that once took minutes now finish within a second. Our enterprise customers' comprehensive recruiting dashboards, which are being filtered and shared in real-time as they make decisions that impact their organization, are their source of truth. Complementing Postgres with ClickHouse enables us to give them a fully reliable experience and handle even larger data as we scale.” - [Elenie Godzaridis](https://www.linkedin.com/in/elenie-godzaridis/), Director of Engineering, [Ashby](https://www.ashbyhq.com/)

>“We first tried to implement Postgres CDC in-house using Debezium but it was way too complex. I knew a managed product built by engineers, whose goal in life is to transform bits from Postgres into ClickHouse, would be better than anything we could do ourselves.” - [Tal Shargal](https://www.linkedin.com/in/tal-shargal-29388671/), Chief Architect at [Seemplicity](https://seemplicity.io/),  [*read the full customer story*](https://clickhouse.com/blog/seemplicity-scaled-real-time-security-analytics-with-postgres-cdc-and-clickhouse)

<iframe width="660" height="371" src="https://www.youtube.com/embed/dSwT5sP1Ryw?si=TAu-vCj6ZrGBU-h2" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>

</br>

![Chart.png](https://cms.clickhouse-dev.com:1337/uploads/Chart_7c2a541127.png)

**This graph shows the growth we witnessed in Postgres CDC to ClickHouse usage over the past year,** [**reference**](https://www.linkedin.com/feed/update/urn:li:activity:7310749593296588800/)**.**

## Product Highlights

<iframe width="660" height="371" src="https://www.youtube.com/embed/s1hREjN02S0?si=J2ByKIkTTU9Mw94m" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>

</br>

The Postgres CDC connector is purpose-built for Postgres and ClickHouse and comes with a vast suite of features. Here are a few noteworthy ones:

1. **10x Faster Initial Loads and Resyncs** – Achieved through [parallel snapshotting](https://blog.peerdb.io/parallelized-initial-load-for-cdc-based-streaming-from-postgres), enabling parallel loading of single large tables. Terabytes of data can now be moved in hours instead of days.
    
2. **Replication Latencies as Low as 10 Seconds** – The replication slot is [continuously consumed without reconnections](https://clickhouse.com/blog/enhancing-postgres-to-clickhouse-replication-using-peerdb#efficiently-flush-the-replication-slot), and ClickHouse's [ReplacingMergeTree](https://clickhouse.com/blog/postgres-to-clickhouse-data-modeling-tips-v2) ensures minimal end-to-end latency.
    
3. **Automatic Replication of Schema Changes**– [Supports](https://clickhouse.com/docs/integrations/clickpipes/postgres/schema-changes) operations like ADD COLUMN and DROP COLUMN with zero manual intervention.
    
4. **Table and Column Exclusions** – Allows fine-grained control for better security (e.g., handling PII) and optimized network throughput.
    
5. **Secure Connectivity** – Supports [**AWS PrivateLink**](https://clickhouse.com/docs/integrations/clickpipes/aws-privatelink) and [**SSH tunneling**](https://clickhouse.com/docs/integrations/clickpipes/postgres#optional-setting-up-ssh-tunneling) for private, secure connections to source Postgres databases.
    
6. **Support for Native Postgres Features** – Includes replication for [partitioned tables,](https://blog.peerdb.io/real-time-change-data-capture-for-postgres-partitioned-tables) TOAST columns, and [advanced data types](https://clickhouse.com/docs/integrations/clickpipes/postgres/faq#how-are-postgres-data-types-mapped-to-clickhouse) such as arrays and JSON.
    
7. **Open API (Beta)** – Enables [programmatic](https://clickhouse.com/docs/integrations/clickpipes/postgres/faq#can-clickpipe-creation-be-automated-or-done-via-api-or-cli) creation and management of pipes. Ideal for ISV applications deploying isolated ClickHouse Cloud + ClickPipes environments. [**Terraform**](https://registry.terraform.io/providers/ClickHouse/clickhouse/3.2.0-alpha1/docs/resources/clickpipe) **support coming soon** for easy infrastructure-as-code automation.
    

## Pricing

As part of the GA launch—and as [mentioned](https://clickhouse.com/blog/postgres-cdc-connector-clickpipes-public-beta#pricing) earlier during the beta—we’re introducing pricing for the Postgres CDC connector. Our goal is to keep it highly competitive while staying true to our core vision: making it seamless and affordable for customers to connect their Postgres databases to ClickHouse. The connector is **over 5x more cost-effective** than external ETL tools and similar features in other database platforms. 

Note that pricing will start being metered in monthly bills **beginning September 1st, 2025 for all customers (existing and new) using Postgres CDC ClickPipes**. Customers will have a 3-month window to optimize costs if needed, though we don’t anticipate most will need to make any changes. For detailed information on the pricing structure, examples, and FAQs, click [here](https://clickhouse.com/docs/cloud/manage/billing/overview#clickpipes-for-postgres-cdc).

## How to try Postgres CDC to ClickHouse

You can follow the links below to connect your Postgres databases to ClickHouse Cloud for blazing fast analytics.

* [Ingesting data from Postgres to ClickHouse (using CDC)](https://clickhouse.com/docs/en/integrations/clickpipes/postgres)
    
* [ClickPipes for Postgres FAQ](https://clickhouse.com/docs/en/integrations/clickpipes/postgres/faq)
    
* [Try ClickHouse Cloud for free](https://clickhouse.com/docs/en/cloud/get-started/cloud-quick-start)

We’re excited to announce the general availability of the Postgres CDC connector in ClickPipes. Natively integrated with ClickHouse Cloud, this connector enables blazing-fast replication of your Postgres databases with just a few clicks.

Postgres CDC connector for ClickPipes is now Generally Available

Today, we’re excited to announce the private preview of the [MySQL](https://www.mysql.com/) Change Data Capture (CDC) connector in [ClickPipes](https://clickhouse.com/cloud/clickpipes)! This enables customers to replicate their MySQL databases to [ClickHouse Cloud](https://clickhouse.cloud/) in just a few clicks and leverage ClickHouse for blazing-fast analytics. You can use this connector for both continuous replication and one-time migration from MySQL, no matter where it's running—whether in the cloud (RDS, Aurora, CloudSQL, Azure, etc.) or on-premises.

The experience is natively integrated into ClickHouse Cloud through ClickPipes, the built-in ingestion service of ClickHouse Cloud. This eliminates the need for external ETL tools, which are often expensive and require significant overhead to set up and manage. This connector is powered by [**PeerDB**](https://github.com/PeerDB-io/peerdb), a fully open source database replication project by ClickHouse—so there’s no lock-in, and the value extends to self-hosted ClickHouse users as well.


[**You can sign up to the private preview by following this link.**](https://clickhouse.com/cloud/clickpipes/mysql-cdc-connector)


After launching our Postgres CDC connector a few months ago—which has been rapidly growing in adoption—there was overwhelming demand from customers to provide similar capabilities for MySQL data sources. We’ve heard all of that feedback and are now launching the MySQL connector in ClickPipes.

<iframe width="768" height="432" src="https://www.youtube.com/embed/9d8EMVHrKkc?si=iSdrgQpS25AnTMKB" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>

## Purpose-built for MySQL and ClickHouse

Our focus at ClickPipes has always been quality over quantity, so any connector we add is built to scale. The same applies to the MySQL CDC connector — it’s designed to handle terabytes of data, delivering blazing-fast performance with replication latency as low as a few seconds. It supports MySQL-native features, including replication of JSON and vector types, along with flexible replication modes, including GTID-based and binary log position-based (POS) options.

## Key Features

The [MySQL CDC connector](https://clickhouse.com/docs/integrations/clickpipes/mysql) comes packed with features designed to make it easy to initiate and manage replication from MySQL to ClickHouse Cloud. Here’s a list of key capabilities:

* **Blazing-Fast Initial Loads (Backfills):** Migrate terabytes of existing data across hundreds of MySQL tables to ClickHouse within a day. You can configure the number of parallel threads to migrate multiple tables simultaneously. Soon, we plan to support intra-table parallelism—using multiple threads to migrate a single large table.  
    
* **Continuous Replication (CDC):** After the initial load, continuously replicate changes—including all DML operations (INSERTs, UPDATEs, DELETEs)—from MySQL to ClickHouse with latencies as **low as a few seconds**.  
    
* **Table and Column-Level Filtering:** Selectively replicate only the tables and columns you need from MySQL—helping reduce data transfer and storage overhead while supporting compliance and privacy needs, such as excluding PII.

 ![tableselector (1).png](https://cms.clickhouse-dev.com:1337/uploads/tableselector_1_ca659eded9.png)
    
* **Schema Changes:** Automatically replicate schema changes during CDC, including DDL commands like ADD COLUMN and DROP COLUMN on MySQL.  
    
* **Flexible Replication Modes:** The connector relies on MySQL’s binary log for replication and supports both file position-based (POS) and GTID-based replication modes.  
    
* **Native Data Type Support:** Fully supports MySQL-native data types—including vectors, unsigned integers, geospatial types, JSON, vectors and more.  
    
* **MySQL Version Compatibility:** Supports all MySQL versions 8.0.1 and above.  
    
* **In-Place Editing:** Modify existing MySQL CDC Pipes to add new tables, adjust sync intervals (data-freshness), or update configuration settings—all without downtime.  

![editsettings (1).png](https://cms.clickhouse-dev.com:1337/uploads/editsettings_1_debffc96f2.png)
    
* **Enterprise-Grade Security:** Securely connect MySQL sources using SSH tunneling, AWS PrivateLink, and IP-based access controls.  
    
* **Built-in metrics and monitoring:** Enjoy a fully managed replication solution with built-in monitoring, metrics, logs, and high availability—eliminating operational overhead for your team.

![metrics (1).png](https://cms.clickhouse-dev.com:1337/uploads/metrics_1_523cd3d942.png)
    

## How to sign up for Private Preview?

You can sign up for the private preview by filling out [the form on this page](https://clickhouse.com/cloud/clickpipes/mysql-cdc-connector). Our team will reach out to you within a few days and work closely with you to provide early access. Given the anticipated demand, there may be a slight delay, but we’ll ensure we connect with you as soon as possible.

The Private Preview entails no cost and is fully free. This is a great opportunity for you to get firsthand experience with the native MySQL integration in ClickHouse Cloud and directly influence the roadmap. Looking forward to having you onboard!

We are excited to announce the availability of the MySQL CDC connector in ClickPipes in private preview. With this, customers can easily replicate their MySQL databases to ClickHouse Cloud with just a few clicks.

Sai Srirampur

Blog posts by Sai Srirampur

Building an enterprise Postgres service in ClickHouse Cloud

Fast, scalable, enterprise-grade Postgres natively integrated with ClickHouse

Postgres CDC in ClickHouse, A year in review

ClickPipes向けPostgres CDCコネクタが一般提供開始

Postgres CDC connector for ClickPipes is now Generally Available

MySQL CDC connector for ClickPipes is now in Private Preview