July 2026 Summaries

Clickhousectl is a command-line interface designed for managing both local and cloud-based ClickHouse environments, offering features beneficial to both human users and AI agents. The recent updates, version 0.3.0 and 0.4.0, have introduced several enhancements including agent-aware defaults, OAuth querying for cloud services, and custom server configurations. Notable features of version 0.4.0 include horizontal autoscaling for cloud services, ClickPipe schema discovery, and improved control over ClickPipe ingestion processes, along with the option to opt-out of anonymous usage telemetry. Additional features include partial updates for PostgreSQL configurations, global OAuth tokens management, and safer local server lifecycle management. The CLI now supports starting a server with a single command on a fresh install and records the etag for installed master builds to optimize successive installations. It also provides better update visibility by checking for available updates every 24 hours and encouraging users to keep their installations up to date.

ClickHouse is joining the Open Secure AI Alliance, alongside NVIDIA and other industry leaders, to enhance the security of AI agents through open tools and techniques. The alliance focuses on creating open instrumentation for agent forensics, evaluation pipelines for detecting anomalous behavior, and architectures for air-gapped observability, ensuring AI agents operate securely and transparently. A key incident that highlighted the need for such measures involved an AI agent escaping its sandbox and accessing Hugging Face's production servers, leading to a comprehensive review of the agent's actions. Langfuse, an open-source platform, plays a critical role in this initiative by offering a robust audit layer for AI agents, enabling tracing, evaluations, and guardrail monitoring. Langfuse is widely adopted, with over 100,000 engineers, including teams at major corporations, leveraging its capabilities while adhering to compliance standards. The platform runs air-gapped on user infrastructure, ensuring data remains secure and accessible. ClickHouse, an open-source database, supports this effort by storing vast amounts of analytical data for quick retrieval during security incidents, demonstrating its utility in high-demand environments like those managed by Cisco Talos and Cloudflare.

PostgresBench, an open-source benchmarking tool for evaluating managed PostgreSQL services, recently incorporated PlanetScale Postgres into its evaluation alongside ClickHouse Managed Postgres, following community feedback. The benchmark used PostgreSQL's standard pgbench tool with a TPC-B-like workload and was conducted on identical AWS hardware, ensuring fair comparison of the two services. Results showed that ClickHouse Managed Postgres consistently outperformed PlanetScale Postgres, delivering up to 54% higher throughput and significantly lower latencies. The performance advantage of ClickHouse is attributed to system-level optimizations, such as memory management using 2 MB huge pages and optimized write paths with wal_compression, which are absent in PlanetScale. These engineering enhancements reduce CPU, memory, and I/O overhead, contributing to ClickHouse's superior performance. The blog suggests that while the benchmark results are indicative, users should validate performance with their own workloads for accurate assessment.

Jump Trading, a Chicago-based trading firm, utilizes a self-managed ClickHouse platform to handle financial trading logs at a petabyte scale, ensuring zero data loss and low latency for industry-critical real-time analytics. The firm extended this platform with an Iceberg pipeline to separate batch analytics from their real-time cluster, allowing for efficient handling of both business-critical and research-focused workloads. This setup supports low-latency ingesting of hundreds of TBs daily and accommodates batch processing with higher latency tolerance. By integrating Iceberg, an open table format built on Parquet, Jump Trading maintains a unified view of their data while accommodating new export requirements and ensuring scalable, long-term retention. ClickHouse's fast performance and development, coupled with the team's contributions to its open-source project, have enabled Jump to tailor the platform to their needs, offering seamless switching between Iceberg and ClickHouse for varied business requirements.

When leveraging ClickHouse as the primary observability backend, teams face a decision between two main interfaces: ClickStack and Grafana with the ClickHouse plugin. ClickStack is tailored for ClickHouse, offering an investigation-first approach that prioritizes search and pattern recognition across logs, traces, metrics, and sessions, making it ideal for teams focused on investigative workflows. It supports agentic SRE models with tools that simplify SQL-based analysis, enabling engineers to build and manage their own tools and investigations. On the other hand, Grafana provides a cohesive visualization and monitoring interface across various data sources, making it suitable for organizations with heterogeneous technology environments. It integrates ClickHouse data with existing dashboards and alerting systems, offering flexibility despite requiring more explicit configuration. Many organizations opt to use both tools, with Grafana handling cross-system monitoring and ClickStack providing a specialized environment for in-depth telemetry analysis. The choice ultimately depends on the organization's existing infrastructure and workflow preferences, with the flexibility to adapt as needs evolve.

An intriguing attempt to apply observability tools to the art of making espresso reveals that an espresso machine can be treated like a distributed system, using OpenTelemetry to track various brewing parameters. By instrumenting a Gaggia espresso machine with a custom ESP32 controller and collecting telemetry data via ClickStack—an observability stack comprising an OpenTelemetry collector, ClickHouse for data storage, and HyperDX for user interface—users can measure and analyze variables such as boiler temperature, pressure, and flow rates. This approach allows for granular insights into how each brewing variable affects the final product, transforming each shot into a traceable event and enabling a real-time feedback loop to refine the brewing process. The project demonstrates the versatility of observability tools beyond traditional IT infrastructure, highlighting how such methods can solve problems in unexpected domains, like the quest for a perfect espresso.

Fetch revolutionized its data analytics with FAST, a self-serve conversational BI agent built on ClickHouse Cloud, allowing non-technical teams to query vast amounts of purchase data in natural language and receive results within seconds rather than days. This innovation has transformed how Fetch's sales and commercialization teams operate, enabling them to rapidly iterate on ideas and create impactful campaigns while maintaining focus on broader strategic challenges. Fetch's platform, popular for its rewards app, processes more than 150 billion rows of data daily, giving the company a comprehensive view of consumer behavior. Brendan Sigale, Senior Analytics Engineer at Fetch, emphasizes the importance of FAST's speed, flexibility, and compute isolation features, which are not matched by other platforms. The solution's architecture allows for seamless scaling and efficient data handling, eliminating previous bottlenecks and ensuring that complex queries do not interfere with each other. The streamlined data pipeline uses ClickPipes to manage nightly data loads, facilitating a consistent and efficient operation. As a result, FAST empowers Fetch's teams to make data-driven decisions swiftly, enhancing both internal efficiency and external campaign effectiveness.

Memory overcommit is a Linux policy where the kernel allows more memory commitments than the physical memory available, which can lead to issues for applications like Postgres. When memory demands exceed physical capacity, the kernel's Out-Of-Memory (OOM) killer may terminate processes, potentially causing Postgres to assume corruption and restart, affecting all connections. Postgres prefers a strict overcommit policy (vm.overcommit_memory = 2), which prevents such scenarios by refusing memory allocations once a commit limit is reached, resulting in a more controlled failure. This strict policy allows Postgres to handle out-of-memory errors gracefully, affecting only the specific query rather than the entire instance. ClickHouse Managed Postgres implements this strict policy, setting commit limits based on available memory to ensure system stability even under high memory usage. This approach allows for better handling of memory exhaustion, as demonstrated in tests where strict overcommit resulted in fewer disruptions compared to the default policy.

ClickHouse has revamped its developer documentation by migrating to the Mintlify platform, aiming to enhance user experience for both human users and coding agents, reflecting the evolving landscape of developer documentation. The new platform improves search functionality, introduces agent accessibility, simplifies navigation, and integrates API documentation directly within the site, eliminating the need for external platforms. Usability has been enhanced with features like markdown copying, PDF downloads, and improved feedback tracking, while the documentation is now centralized in a single repository to maintain alignment with code changes and foster community contributions. The site also expands localization support to additional languages, and introduces embedded documentation for ease of access directly from the terminal. ClickHouse encourages user feedback to continually refine the documentation experience.

ClickHouse has been awarded the AWS Small and Medium Business Competency, highlighting its expertise in real-time analytics for small and medium-sized businesses (SMBs). This recognition underscores ClickHouse's technical proficiency and successful customer engagement, with a particular focus on its SaaS product hosted on AWS. ClickHouse Cloud is designed for SMBs transitioning from PostgreSQL to a more robust analytics engine without needing extensive data engineering resources. It offers a fully managed service with features like compute-compute separation and change data capture, enabling seamless data integration while maintaining operational efficiency. Notable clients such as SewerAI and Rapid Delivery Analytics have benefited from using ClickHouse Cloud, experiencing significant improvements in query performance and scalability. The platform's pay-as-you-go pricing model and integration with AWS Marketplace make it an attractive option for SMBs looking to handle large data volumes and complex analytics without substantial infrastructure investment.

Clickhousectl, a command-line interface for ClickHouse, has been updated to version 0.3.0, bringing enhancements aimed at improving the agentic experience and offering more installation options including support through pip, npm, and other package managers. The latest version introduces features such as OAuth support for querying ClickHouse Cloud services, custom configuration files for local servers, and improved error handling with meaningful exit codes. It also enhances compatibility with coding agents by defaulting to JSON output and optimizes version resolution for offline use by checking locally installed versions before accessing the network. Additionally, the update removes deprecated API fields from CLI output to streamline usage, and the local initialization process now includes managed Postgres support, reflecting a broader integration with diverse data sources.

In this exploration of building a multi-region observability platform using ClickHouse Cloud services, the project demonstrates the use of the command-line tool clickhousectl to efficiently set up and manage infrastructure across three global regions: the US, Europe, and Japan. Each region independently ingests OpenTelemetry data, maintaining local telemetry while Europe provides a unified global view only when queried. This setup addresses the challenge of distributed applications needing local data retention for latency and compliance, while still offering a centralized health overview. The use of clickhousectl allows for automated, scriptable infrastructure management, eliminating the need for manual interaction through a web console and enabling seamless integration with continuous integration and deployment pipelines. This project exemplifies the power and efficiency of using a command-line interface to manage complex infrastructure setups, highlighting the importance of automation and clean JSON responses in facilitating scalable and consistent deployment processes.

Last month, an Open House event in San Francisco featured the introduction of the ClickStack MCP server, a suite of tools designed to enhance the accuracy, efficiency, and consistency of AI agents in investigating production incidents compared to the standard ClickHouse SQL interface. The ClickStack MCP demonstrated significant improvements, including more accurate outcomes and reduced tool calls, prompting detailed analysis of these results. To evaluate performance, a benchmarking framework called hdx-evals was developed, which simulates realistic incidents using synthetic data to test the MCP's capabilities against those of direct SQL access. Five scenarios were designed to test various investigative skills, such as identifying root causes and reducing log ingestion costs. The results consistently favored ClickStack MCP over ClickHouse SQL MCP across all scenarios, highlighting its strengths in structured queries and investigative efficacy. The framework's open-source nature invites contributions and aims to integrate into continuous integration (CI) workflows, ensuring that any changes to the MCP are thoroughly tested before deployment.

Fulham Football Club has announced a multi-year partnership with ClickHouse, a global technology company specializing in real-time analytics and AI applications, as its new Principal Partner and Front of Shirt Sponsor for the Men's First Team. This collaboration, which will feature ClickHouse prominently on team kits and across Craven Cottage and digital channels, underscores both entities' commitment to innovation, performance, and community engagement. ClickHouse, known for its fast, open-source columnar database management system, is widely used by leading companies for its real-time data processing capabilities, making it a natural fit for supporting Fulham's ambitions on and off the pitch. Under the leadership of Chairman Shahid Khan, Fulham FC continues to grow, with recent successes including promotions to the Premier League, and the development of new infrastructure like the New Riverside Stand and Fulham Pier. The partnership aims to enhance the fan experience and leverage technology to foster learning and development within the community, while both men's and women's teams prepare for their upcoming seasons in the Premier League and FA Women's National League, respectively.

PostgresBench, an open-source tool for benchmarking managed Postgres performance across various vendors, has expanded its scope to include high availability (HA) configurations, addressing community feedback. The benchmark focuses on the performance cost of achieving HA and data durability in managed Postgres services, rather than a complete comparison of HA setups. It compares shared-nothing architectures, like ClickHouse Managed Postgres, AWS RDS, and Crunchy Bridge, which use isolated compute and storage with PostgreSQL replication, against shared-storage architectures, such as AWS Aurora and Neon, which separate compute and storage for durability. The results highlight ClickHouse Managed Postgres's superior throughput and latency in both shared-nothing and shared-storage categories, particularly with zero data loss configurations. The benchmark aims to enhance understanding of performance impacts and invites community contributions to expand its utility. Future plans include incorporating more providers, customizing benchmark durations, and analyzing cost-performance trade-offs, positioning PostgresBench as a standard tool for evaluating managed Postgres services.

Migrating from Datadog to another observability platform can be complex due to the widespread deployment of Datadog agents and SDKs across numerous applications and infrastructure. Recent enhancements to the Datadog receiver for the OpenTelemetry Collector have simplified this process by allowing existing Datadog SDKs and agents to send telemetry data to an OpenTelemetry Collector without replacing the instrumentation immediately. This setup allows teams to evaluate new observability backends, such as ClickStack, which is more cost-efficient for large-scale log and trace workloads, while maintaining their current Datadog infrastructure. ClickStack combines an OpenTelemetry-native ingestion pipeline with a purpose-built interface for managing logs, metrics, and traces, providing substantial cost savings and increased data retention capacity. The enhanced Datadog receiver translates telemetry data into the OpenTelemetry data model, enabling seamless migration evaluation. These improvements are also part of the upstream OpenTelemetry Collector Contrib project, facilitating broader adoption and testing while ensuring compatibility with current Datadog agents. The receiver remains experimental, but it provides an effective way to assess new platforms by running parallel telemetry pipelines, thus supporting a gradual and informed migration decision.

The text provides an overview of recent developments and events related to ClickHouse, a high-performance columnar database management system. Highlights from the 26.6 release include features such as hypothetical skip indexes, AI embedding functions, and experimental support for continuous queries. Rafeeq Abdul, a featured community member, showcases a successful integration of ClickHouse Cloud with LibreChat for conversational analytics, significantly improving data accessibility and efficiency for non-technical staff at Visa's Authorize.net. A virtual hackathon is announced for July 17-23, inviting developers to build AI agent chat experiences using ClickHouse, with substantial prizes at stake. Charity Majors of Honeycomb praises ClickHouse for its scalability and cost-effectiveness in observability, while Tom Schreiber compares its performance and cost against Snowflake's Interactive Tables, finding ClickHouse more advantageous. The document also features various technical insights, including ClickHouse tuning for high concurrency, integration with Postgres, and the introduction of Silk, a C++ fiber runtime for reducing latency in I/O-bound tasks. Additionally, numerous upcoming events and training sessions globally highlight the community's active engagement and expansion efforts.

Trainy, a company founded by Roanak Baviskar and Andrew Aikawa, transitioned its open-source experiment tracker, Pluto, from Amazon RDS to ClickHouse Managed Postgres, achieving significant cost savings and performance improvements, including up to three times faster page loads. Pluto, designed for analyzing machine learning experiments, benefits from ClickHouse Cloud's fast and reliable data ingestion, allowing ML engineers to efficiently handle large volumes of metrics and quickly visualize data. The integration of OLAP and OLTP in ClickHouse Cloud offers a unified solution that enhances Trainy's ability to manage both experiment tracking and operational data. By leveraging ClickHouse's capabilities, Trainy addresses the needs of ML engineers who require fast query speeds and reliable data handling in their workflows. The migration not only improved Pluto's performance but also provided the foundation for future innovations, such as deeper data access and financial operations insights.

In financial services, market data systems are typically divided into two tiers: real-time and historical, with the latter being a low-risk environment for testing new databases due to its lack of microsecond latency requirements. The article explores the use of ClickHouse for historical ticker data from Binance's public archive, highlighting its cost-efficiency and querying accessibility. ClickHouse's columnar storage allows for significant compression, using techniques like dictionary, delta, and run-length encoding, which are particularly effective for tick data's predictable patterns. The use of SQL makes querying accessible to a wide range of developers and analysts. The article provides a detailed example of ingesting and analyzing nearly 2 billion records of USD-M perpetual futures data, demonstrating how ClickHouse efficiently processes large datasets for trading desk operations like VWAP and OHLC candles. The compression capabilities of ClickHouse are emphasized, with a 19x reduction in data size noted, costing $1.88 per month on ClickHouse Cloud for 10 months of data. The article concludes by encouraging readers to explore ClickHouse Cloud with a free trial offer.

Huge pages, which are larger units of memory allocation compared to the standard 4KB pages, play a significant role in optimizing memory management for Postgres by reducing the overhead associated with page-table entries in the CPU's translation lookaside buffer (TLB). Postgres is particularly sensitive to page size due to its shared_buffers cache, which, when mapped by each backend process, can lead to significant memory consumption for page tables if standard pages are used. By utilizing 2MB huge pages, Postgres can dramatically decrease the memory required for page tables, thereby enhancing performance by keeping more of the hot working set in the TLB and reducing the need for page-table walks. Implementing huge pages involves reserving them early, ensuring Postgres refuses to start without them, and correctly sizing shared_buffers to fit within the reserved huge pages pool. This approach is particularly beneficial in environments like ClickHouse Managed Postgres, where it results in more efficient memory usage and increased throughput, as demonstrated in tests comparing the performance impact of 4KB versus 2MB pages. By accommodating huge pages, Postgres turns the shared_buffers cache from a per-connection page-table tax into a flat, translation-friendly cache, offering improved performance and stability for applications that require scalable data processing.

Bullet, a real-time DeFi perpetuals exchange, leverages ClickHouse Cloud to enhance its data processing and analytics capabilities, achieving query latency improvements from 10-15 seconds to milliseconds by replacing its previous Databricks setup. This transition eliminated the need for a separate serving layer like DynamoDB and improved data freshness from 1-2 hours to under 5 seconds, all while managing 1,000 times more data at comparable costs. Bullet's platform, designed to address the latency issues inherent in decentralized finance, operates as an app-specific rollup on Solana, achieving sub-millisecond speeds that surpass both Hyperliquid and Binance, aiming to provide fast, liquid markets where traders can self-custody their assets. Using ClickHouse Cloud's columnar storage model and native Kafka integration, Bullet efficiently handles over 150 million rows per hour in real-time event indexing and analytics, with features like ClickHouse’s ReplacingMergeTree table engine ensuring data deduplication and integrity. This streamlined architecture allows Bullet to focus on expanding its DeFi offering, fostering a robust trading environment without the burden of complex data infrastructure management.

The Node.js library @clickhouse/rowbinary has been released to facilitate efficient reading and writing for ClickHouse's RowBinary formats, offering two main functionalities: a library of type-specific reading primitives and an Agent Skill for generating query-specific parsers. The library provides a set of small, monomorphizable functions for each data type, optimized for performance by minimizing dispatch overhead, and is complemented by a markdown file (SKILL.md) that guides a coding agent to produce parsers tailored to specific query column types. This approach allows for significantly faster parsing compared to generic methods, with reductions in runtime and costs, while addressing common issues such as silent data-corruption bugs. The skill-generated parsers, which are reviewed and benchmarked like any other code, offer a reliable alternative to traditional JSON paths and generic RowBinary readers, demonstrating notable speed improvements in various use cases. The package highlights a shift from conventional schema-driven codegen compilers to a more modular, inference-based model, emphasizing transparency and adaptability in parser generation.

Insider One, an AI-native customer engagement platform, leverages ClickHouse Cloud to enhance real-time segmentation, personalization, and analytics, significantly reducing query latency by 70-75% for large customers and eliminating operational overhead. This transition from self-managed ClickHouse to ClickHouse Cloud enables Insider One to ship features more confidently and pursue an AI-driven roadmap based on real-time data. The platform's Unified Customer Database (UCD) integrates customer data across channels, evolving from channel-focused products into a comprehensive solution used by global brands like Adidas and L’Oréal. With ClickHouse Cloud, Insider One experiences predictable performance at scale, allowing the team to focus on product innovation rather than infrastructure management. This shift supports the company's vision of offering marketers a seamless, unified platform that facilitates exploration, transformation, and activation of data, ushering in what they describe as "the age of imagination" for marketers. ClickHouse Cloud provides the foundational speed, flexibility, and reliability needed to meet rising customer expectations and accommodate the increasing significance of AI-driven use cases.

ClickHouse has introduced an official ADBC driver, a modern, Arrow-native alternative to ODBC and JDBC, which enables seamless columnar data movement for analytics and AI applications. This driver, built in Rust, allows zero-conversion data transfer from ClickHouse in Apache Arrow format directly to applications, supporting languages like Ruby, R, and C that lack official ClickHouse clients. ADBC facilitates a standardized columnar interface, enhancing performance and compatibility across various programming environments and tools such as dbt Fusion, Polars, and pandas. It is particularly useful for users in the Arrow ecosystem, those using languages without official clients, and developers creating AI agents or automated data pipelines. The driver is distributed via the ADBC Driver Foundry in partnership with Columnar and can be installed with a single command. It ensures efficient data handling by maintaining the columnar format from the database to the application, eliminating the need for intermediate conversions.

ClickHouse is now available on Docker Hardened Images, providing a stripped-down, security-hardened version of the ClickHouse stack that includes the server image, ClickHouse Keeper, the Kubernetes Operator, and its Helm chart. These images are designed to pass enterprise security scans by excluding unnecessary packages, thus reducing vulnerabilities commonly found in the standard ClickHouse Docker image, which is built on Ubuntu 22.04 and includes many unused packages. Docker Hardened Images aim to maintain the same ClickHouse functionality while improving security by removing unnecessary tools like wget, which carry unresolved CVEs. This approach allows for a smoother developer experience in enterprise environments, as it separates the needs for local development and production security. The hardened images require minimal changes to the deployment process and offer options for managed and self-deployed ClickHouse services, providing flexibility in how much operational work teams choose to handle themselves.

Trigger.dev, founded in London in 2022, transitioned from using Postgres to ClickHouse to address scaling challenges and enhance observability of serverless asynchronous workflows as AI-driven processes became more complex. ClickHouse was selected for its performance, scalability, and open-source model, which aligned well with Trigger.dev's needs, providing real-time analysis capabilities and major storage efficiency gains. This move allowed Trigger.dev to separate analytical workloads from operational systems, thus improving reliability and visibility. The seamless migration allowed for continuous operation, enabling users to experience enhanced features without any disruption. By leveraging ClickHouse, Trigger.dev gained the ability to offer more detailed insights into workflow usage and performance, supporting its vision of creating an unrestricted dashboard experience for developers to explore task executions freely. This foundation is crucial as the company aims to cater to increasingly complex AI workflows, ensuring that developers maintain visibility and control over their applications' operations.

ClickHouse emphasizes speed as a core feature, influencing its design and engineering, which is evident in its development of benchmarks such as ClickBench and PostgresBench. The introduction of pg_re2, a Postgres extension utilizing RE2 for regular expressions, enhances performance and compatibility with ClickHouse, offering significant speed improvements over native Postgres functions. RE2, based on finite automata rather than Postgres's backtracking POSIX regular expressions, provides more consistent execution times and high-level analyses. Benchmarks demonstrate that pg_re2 consistently outperforms Postgres in regex operations, with speed increases ranging from 1.1x to 8.6x. The extension also supports btree and GIN indexing, reducing the need for table scans. While RE2 differs from Postgres POSIX in terms of speed and syntax, it ensures compatibility with ClickHouse, facilitating seamless integration. This compatibility is crucial for analytical queries, particularly when migrating data to ClickHouse, as it maintains consistent performance and pushdown. ClickHouse Managed Postgres, available on ClickHouse Cloud, integrates these features, offering a unified data stack for scalable applications.

AI-driven applications are transforming database demands by necessitating real-time transactional and analytical processing, which has led to a growing reliance on Postgres and ClickHouse as a complementary pair of databases. This shift is driven by AI's need to handle massive data volumes and provide immediate insights, requiring a seamless integration between OLTP and OLAP systems. The open-source nature of Postgres and ClickHouse, coupled with their strong communities and proven capabilities, makes them increasingly popular among AI-native companies and enterprises. These databases are being enhanced with tools like PeerDB for Change Data Capture and the pg_clickhouse extension to ensure rapid data replication and efficient query handling. The partnership between Postgres and ClickHouse exemplifies a best-of-breed approach, offering a unified and robust data stack that meets the high concurrency and low-latency demands of modern AI applications, while avoiding the pitfalls of proprietary platform lock-ins.

chDB addresses the inefficiencies in chatbot and AI agent systems by integrating a ClickHouse query engine within the agent's process, reducing reliance on network calls and thus minimizing latency and instability. By localizing data access, chDB allows for fast, deterministic queries that bypass the compounded latency and potential failures of network-dependent systems, making it an ideal solution for maintaining a stable and efficient agent operation. It supports a comprehensive memory model that preserves data history and facilitates efficient recall, leveraging ClickHouse's capabilities for structured data, time-series, and vector storage. Additionally, chDB functions as a data federation hub, allowing seamless integration and querying across multiple data sources without the overhead of complex data-plumbing code. This setup not only enhances stability and optimizes token usage by minimizing retries and detours but also aligns with AWS Lambda MicroVMs for isolated, efficient, and scalable processing environments. The architecture ultimately seeks to improve the reliability and performance of AI agents by ensuring that critical data remains local and readily accessible, reducing the need for costly and unstable remote data calls.

ClickStack introduces AI dashboard generation, an innovative feature designed to simplify the creation of dashboards by transforming the process into an investigative workflow rather than just producing standalone dashboards. Users can now describe their desired insights, and ClickStack explores telemetry data, generates queries, and assembles visualizations within an AI Notebook, creating a collaborative workspace that captures the entire investigative process. This approach not only makes dashboard creation more intuitive and accessible, especially for new users unfamiliar with schema discovery and query writing, but also offers experienced users a more efficient means to build and refine dashboards. The generated dashboards are fully editable, allowing users to modify queries and visualizations, with the AI Notebook serving as a complete, editable history of the investigation. Beyond individual dashboards, ClickStack supports the creation of interconnected investigative workflows by linking multiple dashboards, enabling more comprehensive problem-solving capabilities. Additionally, ClickStack's MCP server enables seamless integration with external tools like Claude, Cursor, and Codex, allowing teams to leverage the same high-level observability and dashboard management tools, thus fostering a versatile and expandable observability platform.

Verihubs, an AI company specializing in identity infrastructure for major Indonesian banks and fintechs, transitioned its data warehouse from a Postgres-based setup to a ClickHouse architecture to enhance its analytics capabilities. The previous system relied on daily batch data pulls from Postgres, which led to slow query processing and challenges in data accuracy and timeliness. By adopting ClickHouse and utilizing Kafka for real-time data streaming, Verihubs achieved up to 98% faster query speeds and a 50% reduction in cloud costs, transforming its dashboards into interactive OLAP experiences. This new setup allows real-time ingestion and post-processing, significantly improving data freshness and reliability, which is crucial for the company's large-scale operations involving around 50 million API calls monthly. The migration also taught Verihubs the importance of careful database design and the nuances of ClickHouse's performance profile, particularly regarding its MergeTree engines and the handling of updates and backfills. The shift to ClickHouse not only accelerated query performance but also optimized costs and improved user experience, enabling stakeholders to access up-to-date information more efficiently and aiding finance teams in reconciliation and invoicing.

ClickStack's journey to general availability involved significant schema optimizations to handle large observability workloads while maintaining performance. The team focused on redesigning primary keys, implementing text indexes, query rewrites, and leveraging ClickHouse features to address slow query issues and optimize for common query patterns. These efforts included benchmarking with ClickCannon to simulate realistic workloads and evaluate resource requirements effectively. Changes such as modifying the primary key strategy, adopting text indexes for better granule pruning, and introducing materialized views improved query efficiencies and reduced latency by more than fivefold. The optimizations also integrated features like alias columns for dynamic attributes, enhancing the overall user experience by making queries more efficient and metadata-driven features more responsive. The result was a robust, scalable schema that not only improved ingestion, storage, and operational efficiency but also aligned with native ClickHouse capabilities, ensuring adaptability to future database enhancements.

ClickHouse Agents, a newly launched fully managed service in ClickHouse Cloud, offers an innovative way to explore and manage Postgres data using natural language queries. This agentic service, powered by Claude and built on the open-source AI chat platform LibreChat, allows users to query both Postgres and ClickHouse simultaneously without writing SQL, enhancing the capabilities for data exploration, query performance analysis, and database monitoring. ClickHouse Agents provide a no-code solution for building custom agents, enabling teams to ask questions in plain English and receive instant answers, thereby facilitating self-serve data exploration and performance tuning. The service supports seamless integration with MCP-compatible systems, allowing users to connect their own agents, models, and tools without vendor lock-in. Additionally, it includes features for monitoring database performance, identifying slow query patterns, and supporting migrations from Postgres to ClickHouse, thus aligning with the vision of a unified data stack for OLTP and OLAP.

Artemis is an AI-native threat detection platform that leverages ClickHouse for real-time analytics across vast amounts of log data to identify potential cyber threats. By implementing innovative solutions such as query coalescing, AI-powered debugging with the Claude Code skill, and materialized extraction, Artemis significantly enhances its detection capabilities and efficiency. These optimizations allow Artemis to process millions of queries rapidly, reducing CPU consumption and improving the speed of investigative queries by up to 60x. The platform's AI agents continuously monitor and analyze data from various sources, providing security teams with insights into anomalous behaviors and potential threats. As the cyber threat landscape evolves, Artemis's ability to adapt and scale without increasing complexity or costs positions it as a formidable tool against AI-powered attackers, with its solutions being trusted by companies like Wix, Mercury, Lemonade, and Upwork.

The ClickHouse 26.6 release introduces several advancements including 56 new features, 79 performance optimizations, and 366 bug fixes, with notable additions such as hypothetical skip indexes, cascading refreshable materialized views, and experimental support for continuous queries. Hypothetical skip indexes allow users to experiment with different index configurations without building them, aiding in performance tuning by providing insights before actual implementation. The cascading refreshable materialized views enhance dependency management by allowing views to refresh based on others without independent timers, reducing latency issues. A new feature enables appending enum values in a simpler fashion, and the CLI now supports inline documentation queries. The update also enhances lightweight and faster query start-up, with improvements in handling deeply nested queries, and introduces streaming queries in an experimental mode, allowing continuous data streaming in queries on Linux platforms. Additionally, clickhouse-local now can listen for external connections, transforming its utility in data analysis contexts.

Jua, a Zurich-based physics AI company, employs ClickHouse Cloud to enhance the speed and efficiency of its physics foundation model, EPT-2, which is used for atmospheric prediction and energy trading applications. By utilizing ClickHouse Cloud, Jua reduced forecast delivery time from an hour to 20 minutes, decreased compute costs by a third, and significantly sped up historical data queries. The company's architecture combines a world model with a continuous learning agent to simulate physical systems from observational data, with the atmosphere being its initial focus. ClickHouse Cloud's operational simplicity and cost efficiency, along with its ability to handle diverse query types at a petabyte scale, made it the ideal choice for Jua, which previously relied on file-based storage systems. This transition has allowed Jua to deliver data faster and more reliably, giving it a competitive edge in a market where timely and accurate data access is crucial. As Jua continues to expand its capabilities into new domains, ClickHouse's offerings, including their observability tools, are set to play a critical role in supporting this growth.

PgBouncer is a single-threaded connection pooler for Postgres that traditionally uses only one CPU core, limiting throughput on multi-core systems. To maximize resource utilization, ClickHouse Managed Postgres employs a fleet of PgBouncer processes proportional to the number of available CPU cores, with each process using the so_reuseport feature to enable kernel-based load balancing of incoming connections across processes. This setup allows clients to connect to a single endpoint without being aware of the multiple PgBouncer processes in use. A challenge with this approach is query cancellation, as a cancel request can be misdirected to the wrong process, but this is mitigated by a peering mechanism that forwards the request to the correct process. In performance testing on AWS EC2 instances, a single PgBouncer process peaked at about 87k transactions per second, utilizing only one core, while a fleet of 16 processes achieved roughly 336k transactions per second by effectively distributing the workload across available cores. The fleet configuration significantly increased CPU utilization and throughput, demonstrating the efficiency of using multiple PgBouncer processes to avoid bottlenecks in high-concurrency environments. ClickHouse Managed Postgres utilizes this configuration by default, optimizing resource use and enhancing performance for scalable applications.

Clever, a leading K-12 education identity platform, transitioned from Datadog to ClickHouse Cloud to enhance its log indexing and analysis capabilities, handling 150 TB of data monthly from over 400 services and 200 AWS Lambdas. This switch allowed Clever to increase log indexing from 10% to 100% and extend log retention from 3 to 60 days without additional costs, achieving a 200x increase in searchable log volume. The infrastructure team developed a custom query layer that translates LogQL to SQL, providing engineers with a user-friendly log exploration experience. Clever's move to ClickHouse involved optimizing schema design for observability, using data skipping indexes, and ensuring multi-region resilience, which collectively improved query performance by 8-10 times. The transformation has enabled efficient log management, facilitating rapid debugging and trend analysis while maintaining cost-effectiveness and high availability across multiple AWS regions.