clickhouse news, how-tos and comparisions

DeepSeek, a Chinese AI chatbot provider, experienced an infrastructure vulnerability that exposed sensitive data from its AI database. Researchers from Wiz found the issue and disclosed it to DeepSeek, which promptly secured the exposure. The exposed ClickHouse database contained chat histories, API keys, backend operational details, and more. This incident highlights ongoing concerns about data privacy with AI models, as highlighted by ongoing investigations from regulatory bodies in Western countries.

ClickHouse consistently outperformed all other tested data stores in both storage efficiency and query performance for analytical queries. The benchmarks demonstrated that ClickHouse is thousands of times faster than leading JSON data stores like MongoDB, DuckDB, and PostgreSQL, while maintaining optimal compression on disk. This makes it an unparalleled general-purpose JSON data store for use cases where cost efficiency and analytical query performance are critical.

In this blog post, Zane Mayberry discusses the implementation of data isolation in a multitenant environment using ClickHouse row policies. The solution involves creating custom settings to enforce tenant-specific SQL queries without requiring numerous roles and policies for each tenant.

The following FAQ summarizes common questions with respect to new tiers introduced in ClickHouse Cloud starting in January 2025.

This technical overview details the migration from AMD64 to ARM64 architecture on AWS Graviton instances for a data-intensive workload, specifically ClickHouse. The document provides insights into compatibility testing, performance benchmarks, and the overall process of transitioning to an ARM-based infrastructure.

BuzzHouse was developed to address critical gaps in the fuzzing landscape used for ClickHouse. By focusing on generating complex yet correct queries and identifying issues beyond simple crashes, it complements the current suite of tools used to test databases.

The provided menu items appear to be for a website or platform related to Ensemble AI. The options include sections like 'Why We Do', 'What We Do', 'About Us', and links to social media platforms along with legal notices such as Service Terms & Conditions and Privacy Policy.

ClickHouse 24.12 introduces a new cache for primary indexes, which can significantly enhance query performance by reducing disk reads and network traffic in distributed setups. This feature is particularly beneficial for tables with frequent point lookups or join operations on primary key columns.

This document provides an in-depth guide on optimizing Clickhouse queries, covering key techniques such as proper use of PREWHERE, leveraging materialized views, understanding parallel processing, and continuous performance monitoring.

The Medallion architecture with ClickHouse provides a structured approach to managing data pipelines through tiered transformations. This self-contained approach enables efficient ingestion, transformation, and consumption of clean, optimized datasets for analytics and applications.

ClickHouse 24.11 introduced several significant features and improvements including the addition of BFloat16 data type for AI and vector searches, pre-warming of Mark cache, enhancements in HTTP interface error detection, and other optimizations.

This post details the process of building a product analytics solution using ClickHouse at ClickHouse, Inc., covering key aspects such as schema design, common queries, materialized views, and integrating additional datasets. It also provides insights on creating an in-house analytics platform named Galaxy.

This case study describes how a team addressed high ingest latency and low throughput issues in their ClickHouse setup using the Kafka Table Engine. By switching from the ProtobufSingle format to RAW_BLOB and increasing the batch size of Kafka messages, they significantly improved performance while maintaining acceptable CPU usage.

Today, we're excited to announce the private preview of the Postgres Change Data Capture (CDC) connector in ClickPipes! This enables customers to replicate their Postgres databases to ClickHouse Cloud with just a few clicks. The experience is natively integrated into ClickHouse Cloud through ClickPipes and supports both continuous replication and one-time migrations use cases from Postgres.

This post compares the cost and management capabilities of Altinity Cloud at Hetzner with AWS and GCP for managing ClickHouse clusters. It highlights how Hetzner's low-cost cloud services can provide a viable alternative for businesses looking to reduce operational expenses.