June 2026 Summaries

LogHouse, the internal logging platform for ClickHouse Cloud, has significantly expanded, now managing 431 PiB of uncompressed data across 1.59 quadrillion rows, a 23-fold increase over two years. It operates across 30+ regions on three cloud providers, handling data with high efficiency, such as 80 GiB/s and 190 million rows per second at peak. This growth is supported by a geosharding strategy that allows writes to remain local to their region, enabling independent scaling while minimizing cross-region costs. The platform uses Async Inserts to manage small write operations efficiently and a three-level table hierarchy that facilitates low-latency, cross-cloud queries by hiding the complex topology from users. LogHouse's development includes features like Distributed tables for seamless data querying across regions and a robust setup for reliable data delivery, even during outages, by leveraging S3 for persistent buffering. Despite its advancements, LogHouse continues to evolve, focusing on reducing memory consumption, enhancing durability in async inserts, improving telemetry without customer impact, and expanding data types to include more OpenTelemetry traces and metrics.

At a recent Databricks Data and AI Summit, Databricks showcased a benchmark for its new low-latency compute product, Reyden, claiming superior performance over ClickHouse, which allegedly crashed during the test. However, the benchmark's lack of transparency regarding hardware, configuration, and methodology raised concerns about its validity. ClickHouse attempted to replicate the benchmark using available details but found no crash, achieving 15,000 QPS with proper sizing and no optimizations. The situation highlights the importance of open, reproducible benchmarks that provide meaningful insights and facilitate fair competition. Benchmark transparency is essential for ensuring credibility and fostering innovation, as demonstrated by ClickHouse's approach of openly sharing configurations and results for public scrutiny and verification.

In exploring the evolving landscape of designing for language learning models (LLMs) and agents, the text highlights the challenges and considerations of creating user experiences (UX) that cater to non-human users. While traditional software and hardware design has focused on human-centric interfaces, the rapid development of LLMs necessitates a shift toward agent-first design, recognizing agents as the primary users. This involves understanding the nuanced differences between human and agent interactions, especially in text-based interfaces, and addressing issues like friction, which can lead to agents opting out of tasks they find cumbersome. The text underscores the importance of testing and adapting products to ensure agents do not discard them due to inefficiencies, focusing on dimensions such as steps to completion, self-correction, consistency, context efficiency, token efficiency, and the strength of familiarity. This approach is crucial as agents increasingly influence how products are discovered, used, and evaluated, with the text suggesting that while some principles may be effective now, they are subject to change as models evolve.

ClickHouse has been awarded the AWS Cloud Operations Competency in the Monitoring and Observability category, showcasing its expertise in managing large-scale observability data without compromising on performance or cost. The platform's columnar storage and compression techniques allow for the maintenance of full-fidelity telemetry, while its parallel query engine ensures rapid responses even under high concurrency. ClickHouse Cloud is leveraged by companies like Modal, Lovable, and Tekion to manage real-time monitoring, AI-powered debugging, and performance metrics, significantly reducing storage requirements and improving query speeds. Security operations also benefit from ClickHouse's capabilities, with platforms like Exabeam and Harvey employing its services for threat detection and network log processing. Additionally, digital-native platforms, such as Qonto and Sony LIV, use ClickHouse Cloud for operational visibility and cost optimization. Observability vendors like Langfuse and Respan have migrated to ClickHouse Cloud to handle large-scale event ingestion and analytics efficiently. As observability evolves into the agentic era, ClickHouse continues to invest in tools like AI Notebooks and the ClickStack MCP server to enhance incident investigation capabilities for both engineers and AI agents.

Tsvetan Stoychev, an ex-Principal Software Engineer at Akamai Technologies, shares his experience using AI tools like GitHub Copilot, Claude, and ChatGPT for vulnerability research in the ClickHouse codebase, despite not being a seasoned bug bounty hunter. Initially inspired by his manager's encouragement and a colleague's expertise, Stoychev explored AI-assisted methods to uncover vulnerabilities, leading to successful submissions to the ClickHouse bug bounty program. He describes a meticulous process of using AI to generate hypotheses, validate them, and create proofs of concept, emphasizing the importance of manual verification to avoid false positives. Stoychev highlights the evolving landscape of AI tools, the necessity of trusted access, and the benefits of using Python for scripting due to its reliability in AI-generated code. He underscores the importance of persistence and flexibility given the rapid advancements in AI technology, which continually offer new opportunities and challenges in cybersecurity research.

Notion has integrated ClickHouse as a native connection within its Custom Agents, enabling users to connect via OAuth and perform read-only queries in plain language without leaving the Notion platform. This integration aims to facilitate "agentic analytics" by allowing data teams to interact seamlessly with their data through the ClickHouse Remote MCP server, which connects to existing tools within developers' and data teams' workflows. Users can add ClickHouse to Notion Custom Agents with a few clicks and configure which tools the agent can use, with all operations being read-only. The integration supports various analytical tasks, such as listing databases, running SELECT queries, and generating daily observability reports, all within the Notion workspace. It allows for the conversion of plain-language questions into SQL queries, providing real-time analytics in the collaborative environment where teams already work.

Vibe.co, now a part of Walmart, harnesses ClickHouse Cloud to deliver real-time campaign reporting for Connected TV advertisers, transitioning from Postgres due to its limitations with handling massive data volumes and complex OLAP workloads. The migration allowed Vibe.co to scale its data from approximately 100 GB to over 2 TB, maintaining performance and flexibility without significant architectural changes, with over 90% of client reports being served in under 100ms. The decision to choose ClickHouse Cloud over self-hosted solutions was driven by operational simplicity, predictable pricing, and strong team support, aligning with Vibe's philosophy of minimizing complexity for enhanced trust and scalability. Rémi Paulin, a Staff Data Engineer at Vibe, emphasizes the importance of simplicity in architecture, reliability, and the flexibility that allows internal teams to innovate without constantly redesigning data pipelines, highlighting the strategic advantage of having a SQL-centric database for both human and AI-driven analytics. This approach not only ensures high performance and trust among clients but also supports future growth and adaptation within the fast-evolving data infrastructure landscape.

At ClickHouse Cloud, ensuring reliable backups for PostgreSQL involves managing write-ahead log (WAL) archival to maintain data durability and recoverability. While WAL-G, a Go-based tool, has been effective in this role, its garbage-collected nature leads to unpredictable memory usage, which poses challenges in resource-constrained environments. To address this, ClickHouse developed WAL-RUS, an open-source Rust-based tool that offers predictable memory efficiency and compatibility with WAL-G. WAL-RUS utilizes Rust's explicit memory management and a daemonized architecture to minimize unnecessary buffering and maintain stable memory profiles, reducing memory consumption by over 70% compared to WAL-G. Benchmarks indicate that WAL-RUS achieves comparable archival throughput while offering a more stable resource footprint. This makes capacity planning simpler and aligns with the needs of modern, scalable applications. WAL-RUS remains fully compatible with existing WAL-G configurations and archives, simplifying migration for current users, and it introduces support for Postgres 17's WAL summaries for incremental backups. The tool is open-source and aims to become the default backup and WAL archival mechanism for ClickHouse Cloud's managed Postgres offering, inviting community feedback and contributions.

Silk is an advanced C++ library developed to enhance the performance of ClickHouse by providing a stackful-fiber scheduler optimized for high-concurrency I/O tasks. Unlike traditional threads, fibers in Silk facilitate cooperative multitasking, making them ideal for asynchronous I/O, which is crucial as distributed systems face growing bottlenecks. Silk leverages io_uring for efficient asynchronous operations and utilizes a topology-aware work-stealing algorithm to optimize CPU resources without steady-state heap allocation. It showcases significant improvements in tail latency and throughput compared to conventional thread-based and coroutine systems, particularly in network-bound environments like distributed caches. With its Linux-specific design, Silk promises substantial performance gains by minimizing allocator-induced delays and providing precise synchronization primitives. The library's benchmarks have demonstrated its ability to handle millions of input/output operations per second with minimal latency, setting a new standard for asynchronous runtime systems in high-performance computing environments.

The May edition of What's New in ClickStack highlights significant updates to enhance dashboard functionality, including the introduction of clickable dashboard table actions and scoped filters to streamline multi-source dashboard creation. Enhancements to the service map now include latency percentiles and throughput metrics, improving the ability to diagnose service performance issues. Key announcements from the Open House event include the beta launch of AI Notebooks for Managed ClickStack users and the private preview of ClickStack Cloud, a serverless observability platform. Experimental support for PromQL aims to integrate Prometheus-style metrics querying within ClickStack, while improvements to trace and event pattern analysis enhance the usability of the software. Additionally, the update introduces a Browser RUM dashboard template, optimized Y-axis scaling for line charts, and several user interface improvements that facilitate easier navigation and data interpretation across various observability tools.

CostBench is a benchmarking tool designed to evaluate the cost-performance of real-time analytics systems by measuring the entire analytics path, from data ingestion to query execution. Unlike traditional benchmarks that focus on static datasets, CostBench assesses the continuous ingest, data maintenance, and query serving processes, highlighting how systems handle fresh data and maintain query readiness. The benchmark was applied to compare Snowflake and ClickHouse Cloud, revealing that ClickHouse outperformed Snowflake significantly in terms of cost-efficiency and query latency. ClickHouse maintains a continuous freshness model with its incremental materialized views, keeping pre-aggregations in sync with raw data, thereby supporting sub-minute decisioning workloads more effectively. Snowflake's Interactive Tables improved query performance compared to its standard materialized views, but the overall cost and complexity of maintaining freshness with Snowflake remained higher. This comprehensive approach underscores the importance of evaluating real-time analytics systems based on their ability to manage and serve continuously changing data efficiently.

Visa has leveraged ClickHouse Cloud and LibreChat to transform its data analytics capabilities, enabling real-time, natural language querying of vast payments data without the need for complex SQL expertise. By integrating these technologies, Visa has dramatically reduced the time required to generate business insights from multi-day processes to sub-second responses, while maintaining strict security and governance protocols. This shift has allowed Visa to efficiently handle over 40 petabytes of data, improving the speed and accessibility of insights crucial for making informed business decisions. The use of materialized views in ClickHouse has been pivotal, providing up to 70 times faster query performance by pre-computing results as new data arrives. The platform's architecture ensures data security and compliance, with encrypted data pipelines and private connectivity. As a result, Visa has reclaimed significant time savings, enabling teams to focus on strategic initiatives and potentially surfacing millions in revenue risk that can be acted upon promptly.

The introduction of the ClickStack MCP server to ClickHouse Cloud enhances observability by providing external agents and AI workflows with advanced investigative tools previously available only in the open-source ClickStack. This development allows Managed ClickStack users to access comprehensive telemetry data without incurring high costs associated with data sampling and rollups. By leveraging the ClickStack architecture, agents can perform in-depth analyses using stable, high-level semantic tools tailored for observability tasks, such as identifying log patterns and tracing system anomalies, rather than relying solely on raw SQL queries. This approach enables more efficient and consistent investigations, as evidenced by internal benchmarks showing improved performance metrics. The ClickStack MCP server also integrates orchestration capabilities, facilitating collaboration through dashboards and persistent searches, and supports scalability by isolating agent workloads from user-facing processes. This integration is designed to meet the growing demands of agent-driven observability, ensuring that investigations can leverage unsampled, long-term data to maintain high fidelity and context.

Last week's release of pg_clickhouse v0.3.2 brings significant improvements to the interface for querying ClickHouse from Postgres, despite being a minor update. It introduces support for PostgreSQL 19 Beta1, allowing for better optimizations and preparation for the final release. Enhancements in TLS connections now offer more secure and flexible connection options, thanks to contributions from Andrey Borodin. The update also fixes errors in regular expression pushdown logic, improving compatibility between Postgres and ClickHouse. Additionally, it addresses memory consumption issues identified in certain customer queries, particularly in nested-loop joins and unbuffered queries with the HTTP driver, and introduces new CREATE SERVER options like compression to streamline data handling. With these updates, the pg_clickhouse and ClickHouse Cloud combination continues to establish itself as a robust unified data stack for scalable applications.

Spyne, an AI-native automotive retail tech company, enhanced its data pipeline efficiency by transitioning from a complex self-managed Debezium/Kafka setup to ClickPipes and ClickHouse Cloud. This migration eliminated time-consuming manual processes, such as onboarding new tables and resolving schema drift issues, which previously demanded extensive engineering hours and created operational challenges. With ClickPipes, automated schema evolution and a user-friendly interface drastically improved developer velocity and data consistency, allowing for rapid deployment and analysis. However, the team encountered new challenges, such as managing eventual consistency and adapting to a lack of granular recovery control in the managed environment. Despite these challenges, the shift significantly reduced engineering toil, allowing Spyne to focus more on analyzing data rather than maintaining the pipeline, ultimately enhancing stability and speed in their data operations.

The future of observability is shifting towards a decentralized model where thousands of AI agents, tailored to specific organizations, teams, and problem domains, replace the notion of a single proprietary AI agent. These agents are becoming the primary interface for observability, capable of querying telemetry, identifying patterns, and generating hypotheses, thus handling much of the investigative workload traditionally managed by human engineers. This transition emphasizes the need for openness, allowing teams to build agents around existing systems and processes rather than conforming to a vendor's framework. As AI agents become more integral, they increase the demands on underlying systems due to their ability to pursue multiple hypotheses simultaneously, necessitating robust data infrastructure. Despite these advancements, human engineers remain central to the decision-making process, with agents providing context and evidence rather than making autonomous decisions. This collaborative model fosters a shared workspace where investigation artifacts can be preserved, reused, and refined, ultimately creating a comprehensive body of operational knowledge for future reference.

ClickHouse, an open-source analytics database, has been celebrated for a decade of collaborative development, likened to a stained-glass window that harmoniously unites diverse contributions into a single, functioning entity. This metaphor emphasizes the importance of community participation and the preservation of individual contributions within the larger framework, akin to how stained glass pieces are held together to create a meaningful image. Highlighting this collaborative spirit, ClickHouse includes contributor credits within its system, enabling a transparent view of its development history. This approach is exemplified by cBioPortal, a platform for cancer genomics at Memorial Sloan Kettering, which uses ClickHouse to achieve faster data processing, facilitating ongoing research. The community-driven model of ClickHouse is further reinforced by initiatives like the Champions program, which acknowledges and amplifies the contributions of educators and advocates within the community. This enduring collaboration ensures the longevity and adaptability of ClickHouse, demonstrating the real impact of open-source contributions.

Appcues transitioned to ClickHouse Cloud to enhance their real-time analytics capabilities, significantly improving query speeds and reducing costs while handling a massive data load of 1.31 PB and 410 billion events. Previously relying on a tech stack that included Airflow and Snowflake, Appcues faced challenges with latency and costs as they scaled, prompting the move to ClickHouse for its efficiency and ability to handle complex data segmentation quickly. The switch allowed them to cut query times by 90% and ingestion latency by 99%, while also reducing analytics spending by 23%. ClickHouse's features, such as SharedReplacingMergeTree and lightweight deletes, enabled Appcues to manage user profiles and GDPR compliance more effectively, while maintaining high performance and low latency. This migration not only improved Appcues’ existing analytics setup but also paved the way for new features, including AI-powered insights and enhanced funnel reports, ensuring their platform supports real-time product decisions and scales efficiently.

METRO Markets leveraged ClickHouse Cloud to revolutionize their data infrastructure, consolidating all company data into a single warehouse for the first time in its history and achieving faster query performance and lower storage costs compared to BigQuery and Snowflake. This transition followed the limitations of their previous self-hosted Hadoop-based stack, which faced scaling and reliability issues as the company expanded. The adoption of ClickHouse has allowed METRO Markets to provide real-time seller analytics, credit risk modeling, and observability, while also enabling new AI applications, streamlining decision-making processes across their European operations. The implementation of ClickHouse has not only improved data accessibility and reliability, but it has also facilitated a unified view of data for different teams, enhancing strategic and operational efficiency. The company's data team, having previously struggled with maintaining the old infrastructure, now benefits from ClickHouse's managed solution, significantly reducing operational burdens and ensuring consistent data availability and speed. Looking forward, METRO Markets plans to further integrate their streaming infrastructure with ClickHouse to expand real-time processing capabilities, indicating an ongoing evolution of their data platform beyond its initial scope.

Migrating Postgres databases between providers often requires more manual effort than anticipated, due to the complexity of maintaining database objects like indexes and triggers, which are crucial for preserving business logic. The blog highlights the challenges of current migration practices, where most managed Postgres providers lack dedicated tools, leaving users to rely on documentation for tools like pg_dump and pg_restore, which can be cumbersome and error-prone, especially with large datasets. To address these issues, ClickHouse has introduced a fully managed Postgres migration service called Postgres to Postgres ClickPipes, which simplifies operations by automating tasks such as schema dump and restore, while also supporting secure connections and version compatibility. ClickPipes, backed by PeerDB, offers features like parallel initial load and reliable Change Data Capture (CDC) to reduce migration time and enhance reliability. Additionally, the service provides robust error handling and notifications, ensuring a more seamless migration experience.

Marking its tenth anniversary as an open-source project, ClickHouse has evolved from a prototype designed to address web analytics data challenges into a global community-driven platform with over 2,000 contributors. The project recently highlighted several advancements, including performance optimizations in version 26.5, such as faster Top-N queries and enhanced join capabilities, which significantly reduce memory usage and processing time. Community member Vasily Chekalkin made notable contributions to the WebAssembly user-defined functions, and new tools like ClickCannon have been introduced for benchmarking ClickHouse's capabilities. The platform has also expanded with services like ClickHouse Postgres and ClickHouse Agents, showcasing its adaptability and integration with other technologies. Furthermore, ClickHouse is now supported in Jaeger, offering impressive performance in handling spans, and various events and training sessions are being held globally to engage with the growing ClickHouse community.

Real-time analytics, once considered a niche use case, is gaining broader recognition due to its critical role in customer-facing analytics, observability, fraud detection, operational applications, and AI agents, all of which require fresh data and low-latency queries. ClickHouse, a company that has been at the forefront of real-time analytics, emphasizes the importance of openness and clear definitions for real-time analytics databases. They argue that benchmarks should be open, transparent, and reproducible, noting that benchmarks often fail to capture essential production properties like data freshness, continuous ingestion, concurrency, transformation latency, and scalability. Real-time analytics platforms must support scalable ingestion without degrading queries, ensure near-instant data queryability, update transformations incrementally, and provide resource efficiency at scale. The rise in demand for real-time analytics is partly due to agentic workloads, which require a continuous flow of fresh data and high concurrency to function effectively. As the industry evolves, ClickHouse advocates for standardized benchmarks and a shared understanding of the comprehensive requirements for real-time analytics beyond mere latency numbers.

Postgres managed by ClickHouse entered public beta in May, offering a fully managed, NVMe-backed Postgres service integrated with ClickHouse, and has since introduced numerous enhancements. These updates include fine-grained access control, a redesigned console, Terraform provider, Stripe integration, and usage metering, all aiming to position ClickHouse Cloud as a leading data platform for both OLTP and OLAP workloads. The redesigned console provides a consolidated view of service information and usage metrics, while the introduction of fine-grained access control allows for more tailored permissions. Infrastructure management is facilitated through a new Terraform provider, enabling infrastructure as code, and Stripe integration simplifies payment processes for managed Postgres services. Additionally, Postgres to Postgres migrations are automated with ClickPipes, and the billing system now incorporates usage metering with detailed breakdowns for better cost management. The platform also employs dynamic capacity reservations on AWS to address increased demand, ensuring reliable service provisioning. Furthermore, Postgres extensions such as pg_clickhouse, pg_re2, and pg_stat_ch have been optimized for better performance and stability. These developments underscore a commitment to enhancing capabilities and integrating seamlessly into existing workflows.

ClickHouse, an open-source analytical database, has become one of the most widely deployed databases since its release under the Apache 2.0 license in 2016, thanks to contributions from over 2,600 developers. Over the past decade, it has achieved significant milestones, including 48,000 GitHub stars, 800 releases, and hundreds of millions of downloads, driven by ongoing enhancements in join performance, SQL updates, JSON handling, full-text search, and vector search capabilities. The database has introduced numerous features such as lazy column replication, runtime filters, and statistics-based join reordering to improve query execution speed and memory efficiency. ClickHouse has also embraced an open benchmarking methodology, ensuring transparency and reproducibility in performance results, and has expanded its ecosystem with tools like clickhousectl, AI functions, and support for multiple SQL dialects. The community-driven development approach has been key to its evolution, and initiatives like the Community Champions Program continue to foster collaboration and innovation as ClickHouse looks ahead to further advancements.

ClickHouse, an open-source database with a robust ecosystem, has become a cornerstone for analytics by integrating with various open standards and technologies across observability, AI, and data storage. With over 2,600 contributors and hundreds of millions of downloads, ClickHouse supports OpenTelemetry for observability, the Model Context Protocol for AI, and open table formats like Iceberg and Delta Lake, allowing for interoperability and flexibility in data management. Its compatibility with other open-source tools like Postgres and integration with platforms such as Apache Airflow and dbt highlight its adaptability and widespread adoption. The ecosystem also includes projects like ClickStack, an open-source observability stack, and chDB, which allows Python integration, further demonstrating ClickHouse's ability to support diverse use cases. As ClickHouse continues to grow, it maintains its commitment to open-source principles, empowering a community-driven approach to innovation and integration.

Open Electricity, an open platform providing live Australian energy data, has significantly improved its data processing capabilities by migrating from Postgres and TimescaleDB to ClickHouse, achieving a 92x increase in query speed. This transition allows Open Electricity to efficiently manage around 1 billion records, generated every five minutes from Australia's National Electricity Market and the Wholesale Electricity Market. The migration not only reduced hardware requirements but also facilitated a more seamless handling of complex queries, particularly those involving market and generation data. ClickHouse's column-store architecture and specialized engine types, like ReplacingMergeTree, have been instrumental in optimizing data aggregation and query performance. This new setup enables real-time data accessibility through a public API, supporting around 300 companies and institutions. The migration process involved running the old and new systems in parallel, allowing for careful traffic management and ensuring consistent API outputs. Challenges such as deduplication, timezones, and memory management were navigated, highlighting the intricacies of handling large-scale energy data. The combination of Postgres for relational data and ClickHouse for analytical workloads has proven to be highly effective, empowering Open Electricity to deliver detailed insights into Australia's dynamic energy grid.

A new GCP Pub/Sub connector for ClickPipes is now available in Private Preview, allowing users to stream data from GCP Pub/Sub into ClickHouse Cloud in real time with minimal setup. This connector supports common formats such as JSON, Avro, and Protobuf, and integrates with schema registry to handle schema inference and message delivery automatically. Users can manage the connector programmatically via OpenAPI and the ClickHouse Terraform provider, offering features like attribute-based message filtering, flexible seek options, and per-key ordered delivery. The connector is optimized for ClickHouse Cloud's architecture, supporting high-throughput topics with built-in metrics, monitoring, and detailed logs for insight into the pipeline. It eliminates the need for routing through GCS or Dataflow, providing a simpler, fully managed ingestion solution that enhances performance and reduces costs.

ClickHouse has introduced ClickHouse Agents, a full-managed agent analytics service integrated with ClickHouse Cloud, announced at the Open House 2026 in San Francisco. Built on the open-source AI platform LibreChat, ClickHouse Agents allows users to create, configure, and deploy AI agents based on live ClickHouse data without requiring SQL or complex setups. The platform offers features such as a no-code agent builder, a managed chat interface, and sandboxed code interpreters, enabling a seamless and secure AI experience. ClickHouse's approach emphasizes openness, composability, and interoperability, contrasting with proprietary systems that lock users into a single ecosystem. This strategy aims to harness AI's power for real-time analytics without vendor lock-in, offering flexibility for users to integrate their own models and tools or use ClickHouse’s native features. The initiative is part of an overarching effort to build an open data platform for AI, with ongoing developments to deepen cloud integration and enhance user feedback mechanisms.

ClickHouse, an open-source analytical database, was officially released on June 15, 2016, and has since become the most popular of its kind with over 2000 contributors, celebrated for its robust development practices and innovative design. Originating from performance optimization experiments in 2009, ClickHouse was built from scratch to address the challenges of processing large volumes of web analytics data in real-time, where existing solutions like MySQL fell short. The database evolved through various iterations, incorporating unique features like column-oriented data storage and a real-time merge tree for updates, before being open-sourced to fill a niche in the market, enabling widespread adoption across industries. ClickHouse's development emphasizes modularity, transparency, and community involvement, providing a platform for contributors to experiment with data structures and performance optimization, with contributions credited in both the changelog and within the system itself. The journey to open-source was driven by a recognition of similar needs in the tech community, leading to a methodical release process that included comprehensive documentation and support systems, positioning ClickHouse as a valuable tool for database management and a learning resource for C++ development.

Cisco Talos, a renowned cybersecurity threat intelligence team, transitioned their extensive 2 PB, 2-trillion-row dataset from a self-managed OSS ClickHouse to ClickHouse Cloud, resulting in significant operational and cost benefits. This migration, which involved no downtime and maintained a 100% data match, reduced storage costs by approximately 90% and total cost of ownership by 75%, while reclaiming over 300 engineer hours annually for product development. The move to ClickHouse Cloud facilitated a shift from maintaining complex infrastructure to focusing on delivering fast and accurate threat intelligence, crucial for Cisco's security products. The new architecture, leveraging managed services and ClickPipes, eliminated the need for custom code and complex orchestration, enabling a more efficient and simplified system. This transformation underscores the team's commitment to enhancing their threat intelligence capabilities and maintaining customer trust.

In an era of unprecedented data collection, ClickHouse and Datadog have announced a partnership aimed at optimizing the storage and analysis of observability data. With the rise of AI applications and distributed systems, telemetry data has reached new levels, prompting organizations like OpenAI and Shopify to adopt ClickHouse for its cost efficiency and fast query performance. This collaboration introduces new features that allow organizations to route logs directly to ClickHouse via Datadog Observability Pipelines, enabling seamless access and search through Datadog Log Explorer. The integration offers several benefits, including the ability to retain extensive telemetry data for longer periods without compromising on performance or usability. By combining ClickHouse's scalable storage capabilities with Datadog's user-friendly experience, teams can conduct investigations and troubleshooting without altering their workflows. This partnership enhances control over data storage, retention, and access, offering a flexible observability architecture that retains data in its original location, thereby avoiding the operational burden of maintaining duplicate copies.

ClickHouse has been recognized with the AWS Retail Competency for its expertise in real-time analytics for retail, reflecting its ability to help retailers transform operational data into actionable insights. This designation signifies AWS's validation of ClickHouse's technical capabilities and customer success, particularly in managing the high-volume, high-cardinality, and high-concurrency workloads required in retail environments. ClickHouse's architecture, featuring columnar storage, a parallel query engine, and compute-compute separation, allows it to efficiently handle dynamic retail demands such as rapid inventory updates, pricing adjustments, and real-time personalization. Prominent users like Mercado Libre and Picnic benefit from ClickHouse Cloud's capacity to provide swift data insights, with substantial improvements in query processing times and cost efficiency. ClickHouse supports a diverse range of retail applications, from marketplace analytics to dynamic pricing and AI-driven shopping assistants, showcasing its role in the ongoing evolution toward real-time, data-driven retail strategies.

Real-time analytics has become increasingly vital as analytics are integrated into products, demanding low latency and high data freshness to enhance user experiences. Unlike traditional batch processing models, real-time analytics must handle continuous data ingestion and provide rapid query responses. The shift towards real-time analytics is driven by embedding analytics directly into user-facing applications, the need for instantaneous data processing by automated systems, and live data-based decision-making. Many existing platforms struggle with real-time demands due to architectures optimized for batch processing, which often lead to delays in data visibility and inefficient scaling under high concurrent workloads. Platforms like ClickHouse, designed from scratch for real-time analytics, address these challenges by ensuring data is queryable within seconds of being written and by maintaining performance and resource efficiency at scale. This approach allows businesses to provide up-to-date insights and handle significantly larger data volumes with the same resources, as demonstrated by companies like Bullet, which saw dramatic improvements in data latency and system efficiency by transitioning to a real-time analytics platform.

ClickCannon is an open-source benchmarking tool designed for ClickHouse that emerged from a project to benchmark observability workloads and build sizing recommendations for ClickStack. Initially developed to address challenges such as determining the necessary hardware to handle observability workloads and simulating realistic user behavior, it evolved into a comprehensive framework for replaying data, generating high-throughput workloads, and measuring performance. ClickCannon can simulate concurrent users, execute realistic query workloads, and collect detailed performance metrics, making it applicable to various ClickHouse workloads beyond observability. The tool provides fine-grained control over throughput, concurrency, and user behavior, allowing users to benchmark custom insert and query workloads. It supports a wide range of throughput and hardware configurations and can utilize both real and generated data for benchmarking. By decoupling reads and inserts and employing a worker-based concurrency model, ClickCannon achieves significant throughput, allowing for meaningful analysis of schema changes and query workloads. Open-sourced to enable broader use, ClickCannon is designed to offer flexibility in modeling realistic workloads, experimenting with pipeline customization, and exploring performance optimizations, thereby providing valuable insights for resource allocation and system configuration.

Migrating Postgres databases between providers often presents complex challenges, including preserving essential database objects and managing operational intricacies, which can deter users without specialized expertise. Traditional tools like pg_dump and pg_restore, while powerful, require manual effort and expertise, making migrations slow and error-prone, especially for large datasets. ClickHouse's new Postgres-to-Postgres ClickPipes, now in public beta, addresses these issues by offering a fully managed migration service that simplifies and automates the process, reducing downtime and operational complexity. This service provides features such as automated schema migration, secure connections, parallel initial loads, reliable Change Data Capture (CDC), and robust error handling, all while allowing for custom workflows if needed. Powered by PeerDB, ClickPipes ensures efficient and reliable migrations by abstracting much of the operational workload, offering users a streamlined and transparent migration experience with enhanced speed and reliability.

ClickHouse Cloud has introduced a multi-stage distributed execution model that allows a single query to scale across numerous nodes, effectively addressing bottlenecks in large JOIN operations and high cardinality aggregations by repartitioning intermediate data between stages. This advancement has demonstrated up to a 3.4x speedup in JOIN-heavy queries and nearly linear scaling in aggregation processes, achieving a 7.4x speedup with eight nodes compared to one. Despite previous capabilities for scaling queries across multiple nodes through physical sharding and distributed table engines, modern petabyte-scale workloads necessitated a solution beyond these methods, which were limited by their inability to freely repartition intermediate results between execution stages. The new model leverages shared storage and parallel replicas for query scaling within ClickHouse Cloud, and introduces an exchange operator to redistribute data between plan stages, enabling efficient large-scale JOINs without requiring a full right-hand hash table on each node. The multi-stage distributed execution model further improves memory efficiency and allows for consistent query performance by dynamically allocating stages across a cluster, utilizing shared storage for data access, and optimizing memory use through shuffling. Initial benchmarks using the TPC-H standard showed significant performance improvements, highlighting the model's potential for scaling analytical workloads and setting the stage for future optimizations with the development of a cost-based optimizer. Currently, this feature is in an experimental phase, available through ClickHouse Cloud’s private preview program.

CostBench is an open benchmark designed to measure the cost performance of cloud data warehouses in terms of "performance per dollar" rather than just speed, assisting users in selecting systems that offer the highest performance for real-time analytics workloads. While traditional benchmarks focus on query execution speed, CostBench emphasizes the inseparability of speed and cost, highlighting that operational costs can significantly affect system performance comparisons. The benchmark addresses the complexity of comparing cloud platforms due to their varied cost presentation models, such as credits and compute units, by focusing on how much computational resource is needed to complete workloads and the associated costs. CostBench evaluates cost performance on two fronts: read-side, measuring query performance per dollar, and write-side, assessing the efficiency of converting fresh ingest into queryable data. The initial release focuses on read-side performance, comparing major cloud data warehouses like ClickHouse Cloud, Snowflake, Databricks, BigQuery, and Redshift, using anonymized datasets and real-world queries. ClickHouse Cloud stands out as maintaining "fast and low-cost" performance even as data scales, outperforming competitors by a significant margin. CostBench's open and reproducible design allows users to validate claims and explore results, ensuring transparency in cost performance assessments.

Observability teams are increasingly handling vast amounts of data due to the rise of AI applications and distributed systems, prompting organizations like OpenAI and Shopify to use ClickHouse for scalable and cost-effective data storage and analysis. A new partnership between ClickHouse and Datadog addresses the historical tradeoff between data retention and accessibility by enabling organizations to route logs directly to ClickHouse through Datadog Observability Pipelines, allowing for enriched, filtered, and transformed log data to be stored efficiently. This collaboration allows teams to utilize Datadog's familiar interface and investigative tools while benefiting from ClickHouse's scalable storage, facilitating historical data analysis and maintaining full-fidelity data without workflow disruption. With federated log search capabilities, teams can search ClickHouse logs directly from Datadog's Log Explorer, eliminating the need for data duplication and enabling seamless integration between the two platforms. This partnership empowers organizations to store and access significantly more telemetry data with enhanced control over data location, retention period, and accessibility, all while preserving performance and usability.

ClickHouse has been integrated into Stripe Projects, offering a new command-line workflow for developers and AI agents to provision real infrastructure efficiently. This integration allows the provisioning of ClickHouse or Postgres services on ClickHouse Cloud with a single command, facilitating the process by automatically syncing credentials to the environment without any user interface interaction. Stripe Projects provides a seamless CLI experience, where users can initiate projects, select necessary services, and manage credentials directly from the terminal, making it suitable for both human and agent executions. The ClickHouse service, known for its high-performance real-time analytics capabilities, and a fully managed Postgres service, currently in public beta, are available through this workflow, ensuring that both services share credentials and reside in the same account. Once provisioning is complete, credentials are directly written to the environment file in an agent-readable format, allowing ongoing operations to be handled by the ClickHouse CLI, clickhousectl, without manual intervention. This developer preview seeks feedback from teams aiming to streamline agent-assisted workflows and reduce the time from repository setup to application deployment. Users can explore these capabilities with ClickHouse Cloud, receiving $300 in free credits to start.

The partnership between ClickHouse and Datadog addresses the growing demand for scalable and cost-effective observability data management, allowing organizations to efficiently store and analyze large volumes of telemetry data. ClickHouse is favored by companies like OpenAI and Shopify for its fast query performance and economical long-term data retention, while Datadog is renowned for its robust observability platform. This collaboration enables organizations to seamlessly route logs into ClickHouse via Datadog Observability Pipelines, allowing for consistent log structuring and context enrichment before storage. The integration allows teams to search ClickHouse logs directly from the Datadog Log Explorer, maintaining existing workflows and ensuring data accessibility without duplication. By leveraging ClickHouse's storage and compute separation, organizations can retain full-fidelity data and execute queries without operational overhead, combining the user experience of Datadog with ClickHouse's scalable storage capabilities. This partnership facilitates a flexible observability architecture, giving teams control over data retention and access while avoiding compromises on performance or usability. The integration is currently available in preview, with interested users encouraged to request access.

ClickHouse, a prominent player in real-time analytics, data warehousing, and AI/ML, has announced a major expansion of its global go-to-market leadership team, highlighting the appointment of Ed Lenta as Vice President for Asia Pacific and Japan (APJ). This strategic move is part of ClickHouse's broader effort to scale its global operations in response to increasing customer demand, following the earlier appointment of Kevin Egan as Chief Revenue Officer. Lenta, who previously held significant roles at Databricks, Amazon Web Services, and VMware, will focus on accelerating the adoption of ClickHouse Cloud and its open-source database across the APJ region. The company also announced several other key leadership appointments across financial services, public sector, solutions architecture, and global revenue strategy, reinforcing its commitment to enhancing its organizational capabilities. These appointments come at a time of significant growth for ClickHouse, which has recently surpassed $250 million in annual run-rate revenue and expanded its customer base to over 4,000. The company's Open House 2026 user conference showcased its integration into the enterprise stack, featuring speakers from major companies like Visa, Cisco, and Intuit.

Qube Research & Technologies (QRT), a global quantitative investment manager based in London, utilizes ClickHouse Cloud to enhance its data platform that serves researchers and operates a real-time risk and P&L system. The firm's adoption of ClickHouse Cloud aligns with its philosophy of being cloud-native from its inception in 2018, diverging from the traditional on-premise data centers used by hedge funds established in previous decades. ClickHouse Cloud enables QRT to overcome limitations associated with traditional databases, such as the number of writers and readers, by providing a more scalable solution. This platform not only supports the development of trading strategies but also serves as the foundation for QRT's observability infrastructure, consolidating logs, metrics, and traces onto a single backend for streamlined monitoring and management. This integration simplifies maintenance and enhances performance, allowing QRT to efficiently utilize its data resources.

ClickHouse has introduced ClickHouse Agents, a fully managed agentic analytics service in ClickHouse Cloud, designed to enable users to create AI agents without the need for coding. Built on the open-source platform LibreChat, ClickHouse Agents allows for seamless integration with ClickHouse data, providing a native AI experience through a no-code agent builder and a managed chat interface. This service supports a variety of functionalities, including secure code execution, multi-agent workflows, and enterprise-level security features. The platform emphasizes openness and interoperability, distinguishing itself from other cloud data platforms by allowing users to integrate both first-party and third-party AI models and tools. ClickHouse aims to offer a flexible, extensible, and vendor-agnostic solution for real-time analytics, enabling enterprises to harness AI's power without being locked into a single ecosystem. With partnerships like AWS and the acquisition of LibreChat, ClickHouse is committed to advancing its AI capabilities while fostering a community-driven approach to innovation.

Executable User-Defined Functions (UDFs) are now available in public beta on ClickHouse Cloud, allowing users to write functions in Python and upload them as a zip file to be invoked from SQL, similar to built-in functions. This approach streams row data to sandbox processes managed by ClickHouse, supporting real-time anomaly detection, among other use cases. While executable UDFs have existed in self-hosted ClickHouse, the new cloud-based implementation allows users to deploy models without managing their own servers. The deployment process is simplified to a single upload in the Cloud console, where models are executed within a managed sandbox. A demonstration involves using a PyTorch autoencoder to score anomalies in stock trading ticks, with the results embedded into a web app using Next.js. This innovation addresses challenges in co-locating data and machine learning models, reducing the need for separate scoring services or complex integration architectures. The system allows inline execution of models with SQL queries, supporting materialized views and enabling real-time data processing. Additionally, the introduction of network-accessible UDFs in private beta enables external data fetching for enhanced context around detected anomalies, showcasing ClickHouse Cloud's potential for integrating machine learning within SQL workflows seamlessly.

DeepL's integration of ClickHouse as a central data warehouse has been pivotal in enhancing its analytics capabilities across various applications, such as website and app analysis, company metrics provisioning, and technical monitoring. Initially adopted in 2020 to build privacy-conscious analytics, ClickHouse's ease of deployment via a single binary proved advantageous over alternatives like Hadoop, enabling DeepL to swiftly establish a Minimum Viable Product (MVP). The MVP, comprising an API for event transmission, Kafka for message brokering, and Metabase for visualization, demonstrated ClickHouse's ability to efficiently handle significant data volumes and rapid queries. Automation became a focus to streamline operations, leading to the implementation of a single source of truth for events and table schemas using protobuf. This setup allowed the creation of complex events and queries essential for understanding user interactions with DeepL, surpassing the capabilities of tools like Google Analytics while maintaining data control and privacy. Over time, DeepL expanded the system to include more data sources, facilitating a transition to data-driven development. The platform's evolution included scaling from a single-node to a 3-shard, 3-replica cluster, processing about 500 million rows of raw data daily. The enriched data infrastructure enabled the development of additional capabilities such as an experimentation framework for A/B testing and a machine learning infrastructure for personalizing user experiences. These advancements allowed for rapid iterations on frontend and backend changes and supported cultural shifts towards data-centric decision-making within DeepL.

Lovable, a rapidly growing Swedish startup, utilizes ClickHouse Cloud to enhance observability and AI-driven debugging for its dynamic AI-generated applications and workflows, enabling engineers to quickly explore billions of logs through natural language prompts. Within just eight months of its launch, Lovable reached $100 million in subscription revenue, becoming one of the fastest-growing software companies ever. The platform, with over 8 million users, allows anyone to build full-stack applications from scratch using AI-driven technology, leading to daily creation of over 100,000 projects and deployment of around 25,000 apps. To manage the complex, probabilistic system of distributed AI-driven services, Lovable employs an OpenTelemetry pipeline that streams logs and traces into a unified ClickHouse schema, visualized in Grafana, providing real-time insights into system behavior. The ClickHouse MCP server plays a crucial role in debugging by eliminating the need for SQL expertise, while the analytics layer supports millions of deployed apps, tracking performance metrics to handle diverse usage patterns efficiently. Lovable also integrates security scanning using ClickHouse, ensuring deployed apps are secure, and continuously expands its use of ClickHouse due to its speed, reliability, and low operational overhead, aligning with Lovable's mission to democratize software creation and infrastructure.

Canva has successfully leveraged ClickHouse to achieve large-scale real-time observability, processing approximately 3 million spans and logs per second, and has significantly reduced infrastructure costs by 70% through redesigning ingestion and storage processes. By optimizing queries and improving schema design, Canva reduced trace search times from 30 seconds to 2.5 seconds, resulting in a tenfold performance improvement. With 240 million monthly active users, the demand on their infrastructure has rapidly increased, necessitating a reevaluation of their observability approach to scale efficiently. The transition involved moving tracing workloads to ClickHouse, simplifying infrastructure setup, and implementing upstream batch processing, which allowed Canva to achieve a 14-fold compression rate in span data and enhance query performance. The logging infrastructure, reliant on Kinesis and AWS Lambda, was optimized using ClickHouse's asynchronous inserts to handle logs efficiently, while search performance was improved through free-text search capabilities using ngram Bloom filters. A custom logging platform was developed to improve the developer experience, featuring a query gateway with advanced search functionalities. This transformation has resulted in a more efficient, scalable observability platform that enhances troubleshooting and insights while reducing resource usage, supported by ClickHouse's excellent support team.

Laravel has launched Laravel Nightwatch, an observability platform designed to provide developers with real-time insights into application performance. Built on AWS Managed Services and ClickHouse Cloud, the platform handles over a billion events daily while maintaining sub-second query latency, enabling developers to instantly visualize application health. By integrating Amazon Managed Streaming for Apache Kafka (Amazon MSK), AWS Lambda, and ClickHouse Cloud, Nightwatch delivers large-scale, low-latency monitoring, preserving Laravel's simplicity and developer experience. The platform's architecture separates transactional and analytical workloads, using Amazon RDS for PostgreSQL and ClickHouse Cloud respectively, to ensure real-time monitoring and analytics. Nightwatch's initial launch saw significant adoption, processing 500 million events and attracting 5,300 users within 24 hours. The platform plans to expand to more regions, enhance data collection and compliance, and support applications at any scale, demonstrating a scalable, cost-effective global monitoring solution while maintaining a developer-friendly approach.

Lyft transitioned from Apache Druid to ClickHouse Cloud to handle over 450TB of data daily and process hundreds of queries per second with simplified operations, integrating both S3 batch ingestion and Kafka/Kinesis real-time streaming. This shift from a self-managed to a managed service significantly reduced operational burdens and facilitated continuous improvements such as automated access control and schema synchronization elimination. Motivated by the goal of connecting people, Lyft provided over 800 million rides to more than 40 million people in the U.S. and Canada last year. The company relies on fast, accurate insights to enhance services and business decisions, processing vast amounts of data for historical analysis and real-time decision-making. Engineers Jeana Choi and Ritesh Varyani discussed the move to ClickHouse Cloud, highlighting its performance, scalability, and simplified infrastructure. They noted the benefits of reduced learning costs, built-in data deduplication, and lower operational costs. The transition encountered challenges, such as adapting to cloud-specific features and reconfiguring internal systems, but resulted in a more scalable and maintainable system capable of efficiently processing terabyte-scale batch data. Real-time analytics, crucial for daily operations, are powered by a streaming pipeline using Kafka and Kinesis, with Apache Flink as the processing layer. The system eliminates manual synchronization of protobuf schema definitions with ClickHouse, using Java reflection for dynamic deserialization. Despite facing challenges in the migration process, Lyft continues to evolve its data stack, focusing on deeper integration within its data platform to improve scalability and operational efficiency.

GitLab transitioned to ClickHouse for its analytics infrastructure to address the need for sub-second insights at large scales, moving away from using Postgres for both analytics and transactional workloads. This shift was driven by ClickHouse's ability to handle extensive data volumes with remarkable speed, reducing query times for over a billion rows from 30-40 seconds to under a second. ClickHouse's integration supports features like Contribution Analytics and GitLab Duo, allowing real-time tracking of engineering outputs and AI adoption. GitLab's strategic pivot included making ClickHouse the default OLAP database across all deployments, while Postgres continues to handle OLTP tasks. The adoption of ClickHouse Cloud reduced operational burdens, allowing teams to focus on delivering high-performance, scalable insights without being bogged down by infrastructure complexities. This move aligns with GitLab's broader strategy to embed analytics deeply into its platform, enhancing the ability to offer real-time, scalable insights across all customer environments.

Netflix processes an astounding 5 petabytes of log data daily, managing an average of 10.6 million events per second using ClickHouse, which is enhanced through three key optimizations. These include faster fingerprinting of logs using a lexer approach, improved serialization with native protocols and LZ4 compression, and efficient querying by sharding tag maps, reducing query times significantly. This architecture allows logs to be searchable within 20 seconds of generation, enabling real-time debugging. Netflix's logging system combines ClickHouse for hot data and Apache Iceberg for cost-effective long-term storage, ensuring engineers can perform interactive debugging and maintain seamless service for over 300 million subscribers worldwide. The success of these optimizations, which transformed Netflix’s logging system into one of the largest and fastest ClickHouse deployments, is attributed to disciplined engineering rather than clever tricks, emphasizing simplicity to handle the company's massive scale efficiently.

Event Deltas, a feature in ClickStack's HyperDX interface, are designed to streamline root cause analysis by automatically highlighting key differences between fast and slow traces, thus accelerating the identification of performance regressions. Unlike traditional methods that require manual filtering and visualization, Event Deltas dynamically analyze trace data to reveal attribute discrepancies correlated with higher latency, such as deployment versions or specific endpoints. This approach allows for quick, interactive exploration of observability data, aiding Site Reliability Engineers (SREs) in pinpointing issues without relying on pre-trained models or static alerts. Recent enhancements to Event Deltas have added flexibility by allowing users to customize visualizations beyond default settings, thereby uncovering deeper correlations within trace data, such as those related to query complexity or resource usage. By integrating seamlessly into existing workflows, Event Deltas serve as an effective exploratory tool for SREs, facilitating a swift transition from observation to actionable insights.

ClickHouse Cloud offers a robust solution for dynamically scaling infrastructure based on resource demands, allowing users to monitor and manage service level agreements (SLAs) for specific queries to optimize user experience. By tagging queries, users can calculate and track SLAs, using clickhousectl to investigate potential breaches and automate responses. The guide details how to define and measure SLAs, illustrating the importance of specific metrics like query latency and concurrency, and provides strategies for identifying and addressing resource pressure through scaling. Automation tools, such as cron jobs and agents, can be employed to monitor SLAs continuously and suggest appropriate scaling actions based on real-time data. The guide emphasizes the use of clickhousectl for managing ClickHouse services, providing a seamless experience from local development to cloud deployment, while ensuring transparency and accountability through an activity log.

BENOCS employs ClickHouse to optimize and monitor network traffic for telecommunications providers, leveraging ClickHouse's fast indexing and fuzzy matching capabilities to handle large volumes of rapidly moving data. Their Flow Analytics tool requires efficient data processing, enabling users to analyze traffic within specific time frames by quickly skipping irrelevant data, thanks to ClickHouse's MergeTree table design. This approach supports a self-healing push architecture, allowing BENOCS to process and correlate terabytes of diverse data feeds daily, even when data arrives asynchronously or is delayed. ClickHouse's ability to perform fast lookups and handle timestamp gaps efficiently ensures that BENOCS can maintain reliable and stable data analysis despite the challenges of variable data streams.

ClickHouse, a powerful analytical engine used by major companies like Cloudflare and Uber, is hosting its first in-person hackathon, Click-a-thon 2026, in Bengaluru, India, on August 1–2, 2026. This event invites teams of 2 to 4 tech professionals to engage in a 24-hour competition across four tracks: Real-Time Analytics, Observability, Data Warehousing, and Agentic AI and Analytics, with a total prize pool of ₹10,00,000. The hackathon emphasizes the creation of innovative projects using ClickHouse, encouraging participants to leverage open-source technologies like ClickStack, Langfuse, and LibreChat. These tools facilitate the development of real-time systems, AI applications, and comprehensive analytics solutions. Applications for the event are open until June 25, and interested teams are encouraged to review the participant handbook for details on eligibility and event structure, with further inquiries directed to the provided email.

In the data industry, "real-time" analytics entails more than just low-latency queries; it requires handling continuous high-volume data insertion while maintaining query speed and freshness. This form of analytics has gained importance as analytics become integral to user experiences, requiring instantaneous data availability for applications like dashboards and automated decision-making. Traditional batch-oriented data systems struggle with the demands of real-time analytics due to architectural limitations, such as batch-optimized ingestion paths and delayed data visibility post-ingestion. As a result, real-time analytics platforms must be designed from the ground up to handle continuous data ingestion, immediate querying, and efficient resource usage at scale. The text discusses the challenges faced by traditional platforms and outlines the key features necessary for a real-time analytics system, using ClickHouse Cloud as an example of a platform that successfully meets these demands by providing rapid data processing and query capabilities.

ClickHouse Cloud demonstrated superior cost-performance in executing the TPC-H SF100 benchmark, outperforming other major cloud data warehouses like Snowflake, Databricks, BigQuery, and Redshift. Using a single 59-core node, ClickHouse Cloud completed the TPC-H SF100 workload, which comprises 100GB of data and 22 complex JOIN queries, in 19.8 seconds at a compute cost of $0.063, making it the most cost-effective solution. In a scaled-down SF10 test, ClickHouse Cloud ran the workload in 2.9 seconds for less than a cent, further establishing its efficiency. The benchmark results highlight ClickHouse Cloud's competitive edge in both runtime and cost across various configurations, with significant improvements attributed to two years of focused JOIN engineering. These advances have made ClickHouse Cloud 26 times faster in handling JOIN-centric workloads, with ongoing efforts to enhance performance for larger-scale operations like TPC-H SF1000 through multi-stage distributed query execution.

The text introduces Aspire and ClickHouse as key tools for developing and managing distributed services in ASP.NET Core applications. Aspire simplifies the orchestration of services, databases, and containers with a small C# program called an AppHost, offering a web UI for resource management and OpenTelemetry integration for monitoring. ClickHouse, a fast open-source columnar database, is highlighted for its suitability in handling large-scale analytical queries with minimal configuration. The demo described in the text showcases an application using Aspire to manage a ClickHouse container, an API gateway with YARP, backend APIs, and a Blazor dashboard, emphasizing the integration of ClickHouse for efficient request logging and analysis. The setup ensures seamless service discovery, stable external endpoints, and efficient data storage via a materialized view that precomputes statistics for analytical queries. The combination of these tools provides a powerful approach to understanding and managing distributed systems, with scalable potential for production environments by leveraging ClickHouse's analytics capabilities and Aspire's orchestration strengths.

Over the past two years, ClickHouse has significantly enhanced its performance on join-heavy analytical workloads, achieving a 26× speed increase on the TPC-H SF100 benchmark compared to version 22.4. This improvement was accomplished through targeted engineering efforts, focusing on making joins a core strength of the system. In the first year, foundational updates like faster parallel hash joins, smarter planning, and aggressive filter pushdown were implemented, resulting in a 4.4× speedup by version 25.4. The second year introduced further enhancements such as correlated subqueries, lazy column replication, runtime filters, and statistics-based join reordering, which collectively contributed to an additional 6× speed increase. These advancements have allowed ClickHouse to execute complex join queries more efficiently and cost-effectively, enabling it to compete with platforms like Snowflake, Databricks, BigQuery, and Redshift. The company plans to continue optimizing join performance with ongoing developments, including distributed joins to handle even larger workloads.

The text discusses the use of ClickStack, an open-source observability stack designed to provide comprehensive insights into distributed systems through OpenTelemetry instrumentation. By setting up two ASP.NET services with OpenTelemetry, users can trace requests from the Order API to the Payment Service, capturing logs, traces, and metrics that automatically correlate across service boundaries. ClickStack stores telemetry data in ClickHouse, offering fast, real-time analytics with features like service maps, trace waterfalls, and log-to-trace correlation without the need for manual configuration. The demo highlights ClickStack's ability to simplify troubleshooting by showing the full path of requests, including database operations, and demonstrates how to use Docker Compose to set up the infrastructure. It emphasizes the benefits of combining OpenTelemetry with ClickStack, such as ease of use, cost-effectiveness, and the ability to handle large-scale data, while offering tools for alerting and dashboarding to monitor and analyze system performance.

The Open House event brought together the ClickHouse community for three days of workshops, technical sessions, product announcements, and discussions on the future of real-time data. Key highlights included the unveiling of ClickStack Cloud, a fully managed serverless observability platform built on ClickHouse, which simplifies infrastructure management for users by allowing them to send OpenTelemetry data to managed endpoints and explore logs, metrics, and traces via ClickStack UI. Additionally, the Managed ClickStack, designed for teams seeking operational control over their observability stack, was announced for general availability. AI Notebooks entered beta, providing a persistent investigation workspace that helps engineers document, share, and explore alternative hypotheses without losing context. A new ClickStack MCP server was also introduced, enabling integration with external AI systems and agents, offering structured tools for observability workflows that go beyond basic SQL interfaces. These initiatives reflect a broader trend towards more collaborative and programmable observability systems, reducing operational overhead while enhancing investigative capabilities.

ClickStack Cloud is a fully managed observability service built on ClickHouse, designed for teams seeking to leverage ClickHouse's performance, scalability, and cost-efficiency without managing their own observability infrastructure. By directing OpenTelemetry data to a managed OTLP endpoint, teams can immediately explore logs, metrics, and traces via the ClickStack UI, with ingestion, buffering, scaling, and storage handled automatically. The platform, currently in private preview, is developing features like automated schema tuning based on query patterns and dedicated query compute for agentic workloads. ClickStack Cloud aims to simplify the user experience by managing complex backend operations, allowing teams to focus on understanding and operating their production systems. Its architecture, based on separating compute and storage, allows for dynamic responses to workload characteristics, ensuring efficient handling of telemetry data. As the service progresses, it plans to offer specialized query compute for exploratory workloads and intensive analysis, enhancing user control and efficiency.

InMobi, a leader in ad tech, optimized its real-time ad reporting API by transitioning from a serverless warehouse to ClickHouse, drastically reducing P99 latency from over 60 seconds to under 3 seconds and cutting monthly costs by 80% from $40,000 to $8,000. This move enabled the system to handle over 400,000 queries daily on 10TB of data, providing a robust and cost-efficient analytical platform that supports InMobi's growth and offers rapid insights into ad performance across various dimensions like device and region. Originally established in India in 2007, InMobi connects over 30,000 brands and publishers to more than 2 billion people in 150 countries, processing around 250 billion ad requests and 30 billion events daily. The "Project Velocity" initiative, presented by engineers Anand Tirthgirikar and Amber Vaid, demonstrated how ClickHouse improved query performance threefold, achieving a 5:1 compression ratio and reducing scanned rows by 70%, all while maintaining strict SLAs. Challenges such as atomic ingestion, concurrent connection limits, and ensuring zero downtime were addressed by implementing a staging-validation-production pattern, asynchronous connection management, and an automatic fallback system. The success of Project Velocity not only enhanced publisher reporting but also established a reusable analytical framework across the company, allowing InMobi's developers to easily leverage ClickHouse's capabilities without deep technical knowledge.

Mintlify has transitioned from using PostHog to ClickHouse Cloud to enhance the performance and scalability of its analytics platform, which serves millions of developers and thousands of companies. The switch has significantly reduced dashboard loading times from several seconds to under a second, improving user satisfaction and the company's Net Promoter Score by approximately 30%. This transition was driven by the need for a more scalable solution that could support real-time analytics and handle the increasing traffic from both human users and AI agents, which are predicted to constitute 90% of traffic by the end of 2026. ClickHouse Cloud, chosen for its ease of use, strong community support, and cost-effectiveness, allows Mintlify to provide reliable analytics without the need for continuous maintenance, eliminating latency and rate-limit issues faced with PostHog. The move has not only improved customer experience but has also reduced operational costs by about 60% while enabling Mintlify to maintain a scalable infrastructure that can support future growth and the rising use of AI agents.

Cogent Security leverages ClickHouse to deliver billions of security detections in sub-second timeframes, supporting its AI-native vulnerability management platform. Transitioning from Postgres to ClickHouse reduced P90 query latency from 5 seconds to under 1 second for over 100 million rows. Cogent's Chart Agent, using an agentic loop architecture with ClickHouse projections, improved accuracy from 40% to 94%. As AI accelerates software vulnerability exploitation, enterprise security teams struggle with complex, often ineffective point solutions. Cogent aims to keep companies abreast of rapidly evolving threats, emphasizing that handling millions of detections requires significant organizational context and understanding of affected systems. The platform's efficiency has been highlighted by the reduction in average time-to-exploit from 2.3 years in 2018 to just 1.6 days by 2026. By addressing the chronic manpower shortages in security teams, Cogent's infrastructure, built on an efficient data pipeline and ClickHouse-driven serving layer, facilitates swift, accurate responses to vulnerabilities, offering a unified view of the security landscape. The company's Ontology Service ensures consistent data model semantics, aiding the agent's understanding of unique customer environments. The speed of ClickHouse is central to Cogent's success, enabling iterative agent loops and rapid evaluation cycles that enhance tool precision and adaptability, crucial for combating AI-accelerated threats.

Socialpruf leverages a fully managed solution using Postgres and ClickHouse on ClickHouse Cloud to offer real-time social analytics for brands, talent agencies, and sports media companies. Originally operating on Databricks' Neon, Socialpruf faced challenges with network transfer costs and connection stability, prompting a transition to ClickHouse Cloud that enhanced Postgres query performance by up to five times while eliminating connection issues. The platform aggregates performance data from social media platforms like Instagram, TikTok, YouTube, and X, providing real-time dashboards and campaign trackers. The transition to ClickHouse allowed Socialpruf to handle millions of rows in milliseconds, significantly improving the user experience and maintaining the product's value proposition. The switch to ClickHouse's managed Postgres services also addressed stability and cost concerns, offering a seamless analytics experience by integrating both OLTP and OLAP functions on a single platform. The migration process, supported by the ClickHouse team, utilized PeerDB for reliable data transfer. Socialpruf now manages thousands of concurrent database connections using PgBouncer, facilitating efficient real-time analytics while maintaining a robust transactional foundation.

Qonto, a digital bank serving over 600,000 small businesses and freelancers in Europe, has significantly enhanced its observability capabilities by adopting ClickHouse Cloud, allowing comprehensive monitoring across its platform without sampling or cardinality constraints. Transitioning from Grafana Tempo to ClickHouse Cloud enabled Qonto to extend query timeframes from a few hours to two weeks, significantly compress high-cardinality data, and reduce storage costs, while integrating OpenTelemetry for semantics and leveraging AI to democratize incident investigation. This approach not only improved incident response times and accuracy but also allowed Qonto's teams to utilize data more effectively, facilitating a shift from a focus on mere infrastructure management to innovative use-case exploration. The integration of AI-powered tools, like the "incident companion," has empowered all employees, regardless of technical expertise, to efficiently query and analyze data, fostering a culture of continuous learning and improvement.

Nava, a company founded in Brazil in 1995 to assist large organizations with data center infrastructure management, developed a real-time payment monitoring platform using ClickHouse for ELO, one of Brazil's largest card networks, processing 22 million transactions daily. This transition from Elasticsearch to ClickHouse reduced storage from 12 TB to 2 TB and infrastructure costs by 87%, from R$900,000 to R$120,000 annually. The platform supports 300 real-time dashboards updating every five seconds, with end-to-end latency under two seconds. The move was driven by increasing transaction volumes and performance limitations of Elasticsearch, with ClickHouse offering significant improvements in query speed and cost efficiency. Lucas Souza, a data engineer and architect at Nava, championed the transition, leveraging ClickHouse's capabilities to enhance performance and scalability while making data querying more accessible with ANSI SQL support. This shift has allowed ELO to detect and resolve network issues swiftly, preventing potential transaction cost increases and reinforcing Nava's reputation in the financial services sector.

Replo utilizes ClickHouse to deliver real-time analytics for Shopify merchants, addressing the challenges of scaling such systems. Initially, Replo's analytics pipeline was simple, with all events flowing into a single table, but as usage grew, inefficiencies emerged. They transitioned to a more structured approach, grouping events by customer and session, and introduced pre-calculated metrics to reduce query costs. A significant challenge was implementing fractional attribution without causing system slowdowns, which was resolved by focusing on real-time data and limiting recalculations to recent events. The team ultimately arrived at a streamlined architecture that maintains fast query performance and minimal write latency. This journey underscores the importance of simplicity and clear trade-offs in building scalable, real-time analytics systems on ClickHouse.

In a comprehensive benchmark comparison using the TPC-H SF100 workload, ClickHouse Cloud demonstrated competitive performance against major cloud data warehouses like Snowflake, Databricks, BigQuery, and Redshift, achieving the best cost-performance ratio. The tests, which involved executing 22 join-heavy analytical queries over 866 million rows, showed that ClickHouse Cloud's single 59-core node was not only efficient in raw runtime but also exceptionally cost-effective, completing the workload in 19.8 seconds for $0.063. The study highlighted the difficulty of direct performance comparisons due to varied compute configurations among cloud providers but underscored ClickHouse Cloud's advantage in delivering high join performance per dollar. Furthermore, scaling down to SF10, ClickHouse Cloud executed the full workload for less than one cent, showcasing its unrivaled cost-performance efficiency. These results stem from two years of focused engineering improvements aimed at enhancing join operations, making ClickHouse 26 times faster than previous iterations. Looking ahead, ClickHouse intends to scale its capabilities for larger datasets with multi-stage distributed query execution, building on its optimized join performance.

Gala, a blockchain-based platform offering games and media, drastically enhanced its data capabilities and reduced costs by transitioning to ClickHouse, an AWS partner, to handle its expanding data needs. Founded by Zynga co-founder Eric Schiermeyer in 2019, Gala initially struggled with its existing data infrastructure, which could not keep up with the increasing data from its fast-paced games and expanding user base. By choosing ClickHouse for its scalability, performance, and cost-effectiveness, Gala improved data ingestion speed, reduced costs by 30%, and significantly enhanced query performance from minutes to sub-seconds. The move also empowered non-technical staff with data access through Metabase, an open-source BI tool, enabling broader usage of analytics across the company. As a result, Gala not only improved its engineering efficiency and marketing strategies but also expanded its data capacity from 3 TB to 9 TB while maintaining robust support from ClickHouse, which addressed system concerns and facilitated the transition process.

Luzmo, originally known as Cumul.io and founded in Leuven, Belgium, leverages ClickHouse Cloud to deliver rapid embedded analytics for data-driven companies dealing with highly variable, user-defined data. Initially adopting the open-source version of ClickHouse, Luzmo transitioned to ClickHouse Cloud to focus on building analytical capabilities instead of managing infrastructure, thus maintaining high performance, flexibility, and cost-efficiency. This platform enables data-intensive applications to provide users with fast and accessible analytics by integrating powerful visualizations and AI-driven exploratory features directly into applications. Luzmo's clients, ranging from marketing and sales platforms to large telecom operators like Proximus, utilize this capability for everything from visualizing familiar SaaS metrics to location intelligence and monetizing vast operational data. ClickHouse supports a wide range of workloads, including embedded analytics, query and plugin log analysis, and vector search, providing a robust infrastructure that scales with demand while remaining cost-effective. Haroen Vermylen, Luzmo's co-founder and CTO, emphasizes the significant role ClickHouse plays in their operations, attributing its high speed, versatility, and cost-efficiency as key factors in their choice, which has allowed them to meet diverse use cases and engineering challenges with custom solutions.

Padlet, a company focused on creating engaging educational software, has transitioned to using ClickHouse Cloud for real-time analytics, enabling teachers to instantly access student engagement data during classes, rather than relying on delayed reports. With millions of global users and a high volume of event data, Padlet needed a solution that combined fast query performance, manageable costs, and minimal operational overhead. ClickHouse Cloud met these requirements, offering rapid data ingestion and powerful query engines, which allowed Padlet to build a streamlined analytics pipeline within a month. This real-time capability helps teachers visualize engagement metrics seamlessly during lessons, rather than as an afterthought, and supports Padlet's mission to create visually appealing and interactive educational tools. As Padlet plans to expand its analytics capabilities, ClickHouse remains central to providing deeper insights into how new features and AI tools are utilized in classrooms worldwide.

Critical Manufacturing, a global enterprise based in Portugal, leverages ClickHouse to transform vast amounts of manufacturing data into real-time operational insights, crucial for industries like semiconductors and medical devices. By transitioning from SQL Server to ClickHouse Cloud, they achieved scalable, fast data ingestion and real-time dashboards with minimal overhead. This shift supports their enterprise manufacturing execution system (MES) software, which is pivotal for high-tech manufacturing sectors requiring constant real-time data access to avoid costly downtimes. The company faced challenges with its legacy systems, which struggled with increased data volumes and operator demands for immediate insights. Discovering ClickHouse through Reddit, the team quickly realized its potential due to its millisecond query responses and ease of setup, leading to its adoption from proof of concept to a comprehensive platform. ClickHouse now underpins Critical Manufacturing's analytics, handling telemetry data and supporting AI-driven features, with optimizations like ELT processes and automatic data lifecycle management. As they continue migrating workloads, they aim to enhance availability and fault tolerance to meet the demands of an ever-connected manufacturing landscape, ensuring seamless data access and operational continuity.

Respan, previously known as Keywords AI, manages a massive large language model (LLM) observability framework, handling around 50 million events daily via ClickHouse Cloud. Initially using a Postgres-based setup, they transitioned to ClickHouse Cloud to accommodate the growing scale and to address Postgres's limitations in handling high write loads. This new architecture supports high-throughput ingestion and rapid analytics, utilizing incremental materialized views for efficient data aggregation, which ensures quick dashboard performance despite data expansion to billions of rows. Respan's platform offers visibility into performance, evaluation, and prompt management for LLM applications, routing requests across multiple models and providers. The company, which participated in the YC W24 batch, has rapidly grown to handle approximately 1 billion monthly requests. Their system design strategically separates real-time metrics from user metadata to optimize query performance and uses a partitioned MergeTree table structure for efficient data querying. By avoiding costly consistency processing and enabling asynchronous data pipelines, Respan maintains predictable and memory-efficient trace analysis. The simplicity and scalability of ClickHouse's indexing model reduce operational complexity, allowing Respan to scale with increasing usage without frequent architectural overhauls.

Rill is an operational BI-as-code tool designed to unify and analyze massive real-time event data generated by AI systems and cloud costs, processing over 100 billion events daily using ClickHouse as its core architecture. By defining data ingestion and metrics with SQL and YAML, Rill allows for version control through Git and facilitates fast development with AI tools like Cursor, enabling precise conversational analysis. The tool addresses the challenge of understanding interconnected data by integrating operational and financial data to provide comprehensive business insights. Founded by Mike Driscoll and Nishant Bangarwa after their experience with Metamarkets, Rill's architecture keeps analytics close to the database, using dlt for declarative data loading and ClickHouse for high-performance aggregation. The system's architecture ensures that business logic and metrics are consistently applied across dashboards, APIs, and programs, supporting real-time interaction with live data. This approach shifts analytics from retrospective reporting to operational decision-making, treating analytics as an integral part of the business infrastructure.

Cloudflare has prioritized building a resilient infrastructure capable of withstanding inevitable failures and surges in traffic, serving about one-fifth of the world's websites. The company has been operating the open-source version of ClickHouse for nearly a decade, showcasing its exceptional query performance by scanning vast amounts of data in under two seconds even during simulated outages. ClickHouse's appeal lies in its node design that reduces negotiation needs, simple HTTP integration, and minimal scaling adjustments, making it suitable for Cloudflare's extensive operations. Jamie Herre, Cloudflare's Senior Director of Engineering, emphasizes that scaling is an ongoing journey of adaptation to more data, complexity, and failures, rather than a finite process. Demonstrations have proven the system's resilience and responsiveness, maintaining performance stability under extreme conditions without sacrificing speed. ClickHouse's features, like "soft cluster" capabilities and SQL dialect, offer flexibility and efficiency, while its open-source community provides valuable contributions and insights. Herre advises companies to prepare for scaling challenges proactively, as disruptions are inevitable, and the key is readiness when they occur.

Trio, a Brazilian fintech company, leverages ClickHouse as its single source of truth for payment processing, compliance reporting, and both operational and customer dashboards, achieving significant storage reduction and high-speed processing with minimal tuning. Handling over 243 million payments in the first half of 2025 and processing over a billion events daily, Trio has integrated ClickHouse to transform its scattered systems into a cohesive analytics platform, ensuring data accuracy and compliance in a highly regulated financial environment. The architecture employs a custom ETL solution and sliding window techniques using refreshable materialized views and ReplacingMergeTree to manage delayed, duplicated, and unordered events effectively. This setup has enabled Trio to maintain accurate, real-time financial analytics while significantly reducing storage needs and enhancing processing speed, allowing the team to focus more on data analysis rather than system limitations. As they continue to refine their system, Trio plans to incorporate recent ClickHouse advancements, aiming for a more sophisticated, reliable, and easy-to-manage analytics infrastructure.

Lago leverages ClickHouse Cloud to handle real-time usage-based billing and complex monetization for large enterprises, scaling from 10,000 to 1,000,000 events per second. Initially a startup, Lago founders realized the critical nature of billing infrastructure while working at fintech company Qonto, where billing complexities rapidly escalated. Aiming to transform the approach to billing, Lago emphasizes open-source solutions that are transparent and adaptable, contrasting with the inflexible models of many billing vendors. As Lago's platform matured, it attracted larger and more demanding enterprises like PayPal, necessitating a robust data infrastructure to support massive usage events. After evaluating several database solutions, Lago chose ClickHouse for its performance and ease of understanding, integrating it with their existing pipelines using ClickPipes. ClickHouse now underpins Lago's billing engine, supporting high-speed ingestion, analysis, and activity logging. As Lago continues to serve larger organizations, ClickHouse plays a crucial role in maintaining performance and reliability, with future plans to consolidate more workloads onto this platform for streamlined operations.

The ClickHouse 26.5 release introduces a range of enhancements, including 38 new features, 51 performance optimizations, and 224 bug fixes. Notably, this release optimizes the ORDER BY ... LIMIT pushdown through joins, offering up to a 20× speed increase and significantly reduced memory usage. It also introduces a GROUP BY ... LIMIT shortcut, a filesystem table function to run SQL against local file systems, and extends Top-N optimizations to GROUP BY queries. The release welcomes new contributors and highlights the growing community that supports ClickHouse's popularity. Additional features include the url_base setting for easier URL management, a negative LIMIT BY for selecting rows from the end of groups, and multi-path SQL/JSON capabilities. An experimental in-browser ClickHouse client is also introduced, and query caching can now be controlled on a per-subquery basis. These updates aim to enhance query performance, flexibility, and usability in ClickHouse, making it easier to manage and analyze large datasets.

ClickHouse Cloud has introduced executable User-Defined Functions (UDFs) in public beta, allowing users to write Python functions that can be uploaded and executed within SQL queries. This functionality enables seamless integration of machine learning models with data in ClickHouse, eliminating the need for separate scoring services or complex SQL translations. An example application demonstrates the use of a PyTorch autoencoder for anomaly detection in equity trades, where each trade is scored inline with its data ingestion. The model is trained on historical data and the results are processed through a materialized view, with anomaly scores being stored for further analysis. Additionally, the system supports network-accessible UDFs in private beta, which can fetch external data to provide insights into detected anomalies, thus simplifying the integration of external APIs and machine learning models directly within SQL workflows. This approach reduces the complexity typically associated with ML streaming data architectures by embedding Python functionality directly in SQL queries.

ClickHouse Cloud introduces a fully managed Postgres service that leverages local NVMe storage, significantly enhancing transaction speeds by up to ten times, and seamlessly integrates with ClickHouse for real-time analytics. The platform offers native Change Data Capture (CDC) and a unified query layer via pg_clickhouse without additional fees, enabling users to effortlessly sync data and build applications that combine transaction and analytical capabilities. This service aims to simplify the traditionally complex integration of separate systems, providing a scalable and reliable foundation for real-time and AI-native applications. Additionally, ClickHouse Cloud offers competitive pricing with a free trial until mid-2026 and discounted rates during the beta period, catering to a wide range of workloads from development to large-scale production deployments. The service has already attracted numerous companies transitioning from other providers, seeking a robust and integrated data stack for various industries, including cyber security, fintech, and social media.