March 2026 Summaries

With the launch of Bring Your Own Cloud (BYOC) on Google Cloud, organizations can now leverage a fully managed ClickHouse service that maintains data residency within their own cloud accounts. BYOC separates the control plane and data plane, where the control plane is managed by ClickHouse and the data plane operates within the user's Google Cloud environment, ensuring data never leaves their account. This setup allows users to integrate ClickHouse with their existing Google Cloud infrastructure, including BigQuery, Dataflow, and Pub/Sub, while maintaining compliance and utilizing the same IAM and VPC controls. The setup involves a three-step process using Terraform for account and infrastructure provisioning, and service creation tailored to user requirements, including networking and resource allocation. The service includes features such as efficient distributed storage, managed backups, scalable infrastructure, and integrated monitoring tools, with upgrades managed seamlessly in the background. BYOC on Google Cloud is available in all public ClickHouse Cloud Google regions, and additional support is planned for more regions over time.

Padlet has developed an engaging, visual platform for educators to create interactive lessons, aiming to foster curiosity, creativity, and collaboration among students. With a user base spanning nearly all countries, Padlet identified a need for real-time analytics to enhance classroom engagement insights, prompting the adoption of ClickHouse Cloud. This solution addressed the challenge of delivering immediate analytics to teachers, allowing them to see student interactions as they occur, rather than through delayed reports. ClickHouse Cloud was chosen for its fast ingestion speed, query performance, and ease of use, enabling Padlet to manage high event volumes efficiently and affordably without extensive infrastructure maintenance. The implementation of a real-time analytics pipeline, built in about a month, allows teachers to access engagement data seamlessly, enhancing the educational experience. As Padlet continues to innovate with new features and AI tools, ClickHouse remains integral to providing deeper insights into student interactions, aligning with Padlet's mission to support a good education through technology.

AI agents are transforming the way databases are used by transitioning from traditional human-centered workflows to AI-driven processes that require real-time data ingestion, rapid query execution, and scalable economic models. ClickHouse, a database platform, aligns naturally with these new AI workloads due to its fast aggregation and support for high concurrency. The development of the Agentic Data Stack exemplifies this shift, integrating tools like LibreChat, ClickHouse, and Langfuse to create a conversational analytics environment that supports AI agents. ClickHouse has introduced various tools and integrations, such as the ClickHouse Assistant and ClickHouse Agent Skills, to enhance the capabilities of AI agents in handling database queries efficiently. The platform's ability to manage large-scale data workloads has attracted companies like OpenAI, Anthropic, and Character.AI, which leverage ClickHouse for its reliability and performance in data-heavy AI applications. As AI continues to redefine database market expectations, ClickHouse aims to be at the forefront by providing a robust infrastructure that meets the demanding requirements of AI-driven data analytics.

ClickHouse has become compatible with data lakes by supporting open table formats like Apache Iceberg and Delta Lake, allowing users to query data directly from these formats without moving it. This development required two years of engineering efforts focused on optimizing query processes, integrating with various catalog systems, and enhancing support for underlying storage formats like Parquet. ClickHouse's new capabilities include processing Parquet files efficiently, working with different open table formats, and integrating with cloud storage solutions such as AWS Glue, Azure Blob Storage, and more. By loading data into its native MergeTree storage engine, ClickHouse offers sub-second query performance, enhancing real-time analytics capabilities while maintaining interoperability with other tools in the ecosystem. This advancement allows teams to leverage ClickHouse for both exploratory data analysis and high-performance queries without sacrificing the benefits of open data storage, providing a versatile and efficient solution for modern data infrastructure needs.

ClickHouse is an open-source columnar database management system optimized for real-time analytical queries on large datasets, making it popular for analytics platforms and data warehouses due to its ability to aggregate billions of rows in milliseconds. Utilizing a columnar storage format, aggressive compression, and vectorized query execution, it requires a deep understanding of its architecture for optimal performance. Proper schema design, efficient data types, and an understanding of partitioning strategies are critical to harnessing its full potential, as poorly designed schemas or inefficient queries can greatly affect performance. The document emphasizes the importance of aligning data structures and query patterns with ClickHouse's architecture to achieve significant improvements in performance and efficiency, such as using the right primary keys, optimizing data scans with skipping indexes, and leveraging materialized views and projections. It also discusses the benefits of using dictionaries for static or slowly changing lookups to enhance query performance, and the text touches on various strategies for data ingestion and management, including the use of system tables for observability and debugging. The document concludes with the idea that while ClickHouse performs exceptionally well out of the box, a thorough understanding of its data management capabilities can lead to substantial gains in storage efficiency and query speed.

The text discusses the integration of Serilog with ClickHouse to improve logging infrastructure in .NET applications, highlighting the challenges of scaling traditional logging systems like Seq or ELK due to cost and performance issues. ClickHouse, a column-oriented analytical database, offers a solution by storing massive volumes of structured logs, which can be queried using SQL, providing fast and efficient data retrieval. The article provides a step-by-step guide on setting up an ASP.NET service that writes structured logs to ClickHouse using Serilog, focusing on schema design, log enrichment, SQL querying, and diagnostics. It emphasizes ClickHouse's advantages, such as 10–20x data compression, full-text search, and cost-effectiveness compared to traditional logging platforms. Additionally, the text illustrates the use of Docker for infrastructure setup and offers insights into optimizing ClickHouse performance for production environments, including batching, partitioning, and real-time alerting through materialized views.

ClickHouse is enhancing its cloud platform to provide administrators with improved tools for optimizing database performance and gaining deeper insights into server behavior. The updates include additional monitoring dashboards and a new Overview page that consolidates key deployment signals for quick health checks. An Infrastructure page offers detailed scaling visibility, featuring metrics on CPU and memory utilization and explanations of automatic scaling behaviors, aiding in better decision-making regarding custom scaling configurations. New notification features alert administrators to potential performance risks, such as merge pressure, failed mutations, and query concurrency issues, with options to receive alerts via email or Slack. These enhancements aim to minimize troubleshooting time, allowing more focus on performance optimization, and are now available in ClickHouse Cloud, which offers a $300 free credit for new users interested in exploring its capabilities.

Cogent Security leverages ClickHouse to quickly process billions of security findings, enhancing their AI-native vulnerability management platform. Transitioning from Postgres to ClickHouse significantly reduced query latency, allowing for rapid data handling across large datasets. This infrastructure supports an agentic loop architecture that boosts accuracy in vulnerability assessment, addressing challenges faced by understaffed security teams overwhelmed by millions of findings. With the rise of AI, the demand for swift and accurate vulnerability management has intensified, as the time-to-exploit vulnerabilities has drastically shortened. Cogent's platform, built on ClickHouse, provides a scalable solution, enabling security teams to filter, contextualize, and remediate risks efficiently. The architecture combines a Knowledge Graph, dbt projects, and ClickHouse's robust querying capabilities to deliver fast, accurate insights. This speed is crucial for security tools to function effectively under heavy data loads, ensuring that vulnerability management keeps pace with AI-driven threats. The incorporation of a conversational reporting system further enhances usability, allowing users to interact with data naturally and swiftly. By implementing an agentic loop with dynamic data exploration and query execution, Cogent has achieved high accuracy in data processing, supported by an innovative Ontology Service that ensures consistent data understanding across varied customer environments. This approach fosters a virtuous cycle where improved speed and accuracy drive more sophisticated vulnerability management capabilities.

The ClickHouse engineering team has reimagined the full-text indexing system to optimize its performance for use with object storage, focusing on sequential access patterns to overcome the latency challenges inherent in remote storage. This redesign allows queries to be executed directly from the index, reducing the need to access the full dataset and thus improving efficiency. The text index is composed of three main components: a dictionary file that stores indexed tokens, a sparse dictionary index for fast lookups, and a posting list that maps tokens to their respective row positions. By employing techniques like block-based layouts, front-coded compression, and Roaring Bitmaps, the new index design ensures efficient storage and retrieval, even when handling vast datasets. Additionally, the index supports complex queries through direct read modes and optimized execution paths, minimizing unnecessary I/O operations. This enhancement aligns with ClickHouse Cloud's architecture, facilitating distributed processing across multiple nodes and leveraging shared object storage for scalable full-text search capabilities.

The text discusses an optimization strategy for auto-scaling database resources, focusing on a new two-window recommender system that enhances both responsiveness and stability in scaling decisions. The original system used a 30-hour lookback window, which led to slow scale-downs and increased infrastructure costs. The new approach introduces a dual-window system with a smaller 3-hour window for quick scale-downs and a larger 30-hour window for stable scale-ups, accompanied by a target-tracking CPU recommendation system to address the limitations of the previous fixed-factor algorithm. This method improves scale-down latency from 30 hours to 3 hours, minimizes oscillations, and reduces costs while maintaining system stability. Additionally, memory-based recommendations and an automatic idling feature further optimize resources during periods of inactivity. Overall, these advancements in the ClickHouse auto-scaling system enhance efficiency and reliability for dynamic workloads, allowing for better alignment of resource allocation with actual utilization.

The March 2026 ClickHouse newsletter highlights several developments, including the introduction of chDB 4, a Pandas-compatible DataStore API that operates on ClickHouse's engine, and the general availability of full-text search that utilizes inverted indexes for efficient token-based filtering. The newsletter showcases the work of community member Jamie Herre from Cloudflare, who presented strategies for managing both growth and failures in analytics infrastructure. Additionally, the newsletter covers Hookdeck's improvements in payload search speed, reducing query latency to under 400ms by implementing typed bucket columns and iterative time-window scanning. The Agentic Data Stack is introduced, offering a system where AI agents can query ClickHouse using natural language, and ClickHouse TTL strategies are discussed for efficient data retention and disk optimization. Upcoming events and training sessions across the globe are also detailed, providing opportunities for users to engage with the ClickHouse community and enhance their skills.

ClickHouse, renowned for its speed in handling extensive telemetry workloads, is primarily favored for observability tasks by companies like Netflix and OpenAI. However, optimal observability performance is achieved not only through ClickHouse's inherent speed but by leveraging query optimization techniques that align with its architecture. ClickStack enhances this by integrating directly with ClickHouse, ensuring queries are crafted for efficiency through methods such as breaking complex queries into stages and utilizing materialized views. These optimizations, including progressive time window pagination, chunked queries, and automatic use of indices, reduce data read, CPU, and memory usage, ultimately accelerating query results. The platform also employs intelligent sampling to maintain performance while analyzing vast datasets and adjusts its settings to exploit ClickHouse's latest features. Beyond its UI, ClickStack aims to make these optimizations accessible through APIs, allowing developers to build high-performance observability solutions without needing extensive SQL knowledge.

AI Notebooks for Managed ClickStack, launched in private preview, introduce a collaborative AI-enhanced workspace for system investigations powered by ClickHouse. The tool integrates AI into the SRE workflow, using Anthropic's Claude models to assist engineers in exploring logs, metrics, and traces within a structured notebook environment. This approach maintains the engineer's control over the investigative process, allowing them to explore anomalies, inspect AI reasoning, and modify ClickHouse queries while preserving context through a visible sequence of steps. Notebooks are designed to enhance, not replace, human oversight, supporting a hybrid workflow where AI proposes actions, and engineers refine and decide. With features like branching workflows and a deep integration with ClickStack's investigative primitives, Notebooks offer a transparent, team-oriented space for complex analysis, encouraging collaborative efforts while ensuring the AI system accelerates rather than substitutes the engineer's expertise. The tool is being gradually rolled out, with plans for further development, including Slack integration, to enhance usability and collaboration.

Adopting AI requires re-evaluating data strategies, as AI workloads demand high concurrency, real-time query processing, and full-fidelity data, which legacy batch-oriented architectures cannot accommodate. ClickHouse is evolving into a unified platform for AI workloads, integrating with systems like Postgres to handle both transactional and analytical tasks. This shift is driven by the convergence of previously siloed use cases, such as data warehousing and observability, and the emergence of AI-driven applications that require a tight integration between transactional and analytical capabilities. Real-time analytics, data warehousing, and observability are converging, demanding platforms capable of handling high-concurrency, low-latency queries for both internal and external workloads. As AI agents become the primary interface to data, data platforms need to provide agent-ready interfaces and observability tools to monitor AI behavior. ClickHouse aims to be the unified data platform for these interactive AI-driven applications, offering a seamless experience that integrates transactional and analytical workloads, modern data warehousing, and AI-SRE observability workflows.

Socialpruf, a Toronto-based platform, utilizes Postgres and ClickHouse Cloud to provide real-time social analytics for brands, talent agencies, and sports media companies. The platform aggregates performance data from social media platforms like Instagram, TikTok, YouTube, and X into dynamic dashboards and reports. Initially using Neon, a Databricks company, Socialpruf faced challenges with connectivity and escalating network transfer costs. Migrating to ClickHouse's managed Postgres service significantly improved query performance and stability, eliminating connectivity issues and reducing operational costs. This migration, aided by ClickHouse's team, allowed Socialpruf to efficiently handle large volumes of data and connections, leveraging PgBouncer for management. Socialpruf's infrastructure, combining OLTP and OLAP, enhances their product's user experience by providing near-instantaneous data processing and visualization, ensuring the company can focus on growth and innovation in the creator economy.

clickhouse-connect is the official Python client for ClickHouse, designed to handle high-volume, performance-sensitive workloads. Initially launched as a synchronous client in September 2022, it gained popularity for its robust feature set. Due to demand for asynchronous capabilities, a workaround using a thread pool executor was employed, although it had limitations like thread pool exhaustion and increased memory usage. To address these issues, a native async client was developed, utilizing the "half-sync/half-async" concurrency pattern. This new design separates asynchronous I/O from synchronous processing, using an aiohttp library for high throughput and a bounded queue to manage backpressure, significantly improving query performance and stability under load. Benchmark tests show that the async client offers a 1.16x speedup on average over the legacy client, with more predictable tail latency and efficient resource usage. The async client is now available in version 0.12.0rc1, with feedback actively sought from users to further refine its capabilities.

The ClickHouse Winter Release introduces 25 new features, 43 performance optimizations, and 183 bug fixes, enhancing its data processing capabilities. Key updates include the production-ready status of the text-index and QBit data type, improved performance for joins, JSON parsing, and insert operations, as well as the ability to batch "infinite" inserts by time. The release also features the embedded ClickStack UI for enhanced observability, faster RIGHT and FULL OUTER JOIN operations, and automatic enabling of minmax indexes for temporal columns to improve query efficiency. The release highlights contributions from new community members and showcases performance improvements through parallel processing and optimized algorithms, significantly speeding up data handling and query execution.

ClickHouse Cloud now supports ingesting data from Google Cloud Storage (GCS) in any order using ClickPipes, enhancing flexibility and efficiency for event-driven analytics. Previously, files needed to be ingested in lexicographical order, but the new unordered mode, aided by Google Cloud Pub/Sub notifications, allows files to be processed as they arrive, regardless of order. This update addresses challenges like backfills and late-arriving data, ensuring that even files that don't follow a strict naming sequence are incorporated into analytics workflows. The unordered mode requires configuring Pub/Sub notifications for OBJECT_FINALIZE events and setting up a service account with the necessary permissions. With exactly-once processing guarantees, the new mode prevents duplicates, and failures are managed through automatic reprocessing. ClickHouse Cloud supports this feature through its Terraform provider, allowing for configuration as code, and plans to extend unordered ingestion capabilities to Azure Blob Storage, complementing existing support for Amazon S3.

The post explores the use of ClickHouse functions for querying and filtering dates and datetimes, focusing on practical applications such as analyzing the New York City taxi dataset. It demonstrates how to set up the database and import over 3 million records from a public dataset. The analysis includes examining taxi trips by hour and 15-minute intervals to identify patterns during rush hours, utilizing functions like `toStartOfHour`, `toStartOfFifteenMinutes`, and `dateDiff` to calculate durations and average speeds. It also investigates the differences between weekday and weekend traffic patterns, showing a sharper spike in weekday trips and a steadier increase during weekends. The post provides a comprehensive look at how these functions can be applied to real-world data analysis, offering insights into traffic flow in New York City.

The February edition of "What’s New in ClickStack" highlights significant updates aimed at enhancing the platform's observability features, offering new query workflows, improved metrics exploration, performance optimizations, and expanded alerting options. A notable development is the integration of ClickStack with ClickHouse Cloud's OpenAPI, allowing users to manage resources like dashboards and alerts directly through the Cloud API for consistent configuration across environments. The update also introduces an embedded ClickStack UI in ClickHouse, a feature designed for local development and experimentation. Users can now toggle between builder and SQL modes in table visualizations, enabling more advanced analytical patterns with raw SQL for comprehensive data manipulation. Enhancements in metrics exploration, such as the new attributes panel, streamline the process of working with OpenTelemetry metrics. The release also brings new visualization options like pie charts and optimizations to accelerate query performance, including skip indices for disjunctions and lazy materialization, which reduce unnecessary data reads and improve efficiency. Additionally, the update introduces alerts on number charts for static threshold-based alerting, contributing to more flexible and powerful alerting capabilities.

ClickHouse offers a comprehensive suite of functions to handle various date formats, such as Unix timestamps, numeric representations, and ISO 8601 strings, making it a versatile tool for data parsing. The fromUnixTimestamp and fromUnixTimestamp64 functions enable conversion of timestamps into DateTime types with different levels of precision, while functions like YYYYMMDDToDate and YYYYMMDDhhmmssToDateTime handle dates encoded as plain numbers. For known format strings, parseDateTime and parseDateTimeInJodaSyntax functions are available, and the parseDateTimeBestEffort family provides flexibility for unknown or mixed formats, with options to handle invalid dates gracefully by returning NULL or defaulting to January 1st, 1970. ClickHouse also supports casting string values directly to date types using the cast_string_to_date_time_mode setting, which can be adjusted for broader format support. This powerful functionality allows users to efficiently parse and manage diverse date formats in their data sets.

Critical Manufacturing, a Portugal-based company specializing in enterprise Manufacturing Execution System (MES) software, transitioned from Microsoft SQL Server to ClickHouse Cloud to enhance real-time analytics and operational intelligence. This shift was driven by the need to handle billions of shop-floor events rapidly and efficiently, offering sub-second insights crucial for high-tech manufacturers across various industries. By adopting ClickHouse, the company significantly improved data ingestion, storage, and querying capabilities, making it possible to manage a vast scale of data and deliver instant insights. The transition allowed the company to utilize ELT processes, ReplacingMergeTree for handling duplicates, and denormalization to optimize performance, while also leveraging ClickHouse for observability and AI workloads. The integration of ClickHouse has transformed Critical Manufacturing's data infrastructure, enabling scalable, cost-effective analytics and supporting continuous operations in a data-driven manufacturing environment. Looking ahead, the company aims to enhance the system's availability and resiliency with plans for multi-replica deployments and sharding as they continue to move workloads from legacy systems.

chDB 4, launched in collaboration with Hex, integrates a Pandas-like DataStore API into the Hex platform, allowing users to write familiar Pandas code that runs on the efficient ClickHouse engine without needing local installations or additional infrastructure. This innovation bridges the gap between SQL and function chain approaches, offering a seamless experience for data practitioners by enabling Pythonic function chaining with the power of ClickHouse's optimized SQL execution. The new DataStore component employs a four-layer architecture that transforms Pandas-style code into ClickHouse execution, utilizing lazy execution to optimize performance and minimize memory overhead. By maintaining Pandas compatibility while leveraging ClickHouse's computational efficiency, chDB 4 addresses common challenges in data science workflows, such as scale limitations and intermediate data handling, and offers smart caching, segment execution, and unified data source support to enhance productivity. This integration is particularly beneficial for users of Hex notebooks, facilitating faster and more efficient data analysis without altering existing workflows, and is complemented by a strong focus on AI-assisted development and Pandas compatibility to ensure seamless integration with modern LLMs.

ClickHouse has introduced Full-text Search as a generally available feature, offering a significant performance boost for token-based text searches by employing native inverted indexes, similar to those in search technologies like Lucene. This new capability enables fast, scalable searches across large datasets, delivering query speeds up to 7-10 times faster compared to traditional methods, and proving more effective than Bloom filters due to its deterministic results and scalability. Full-text Search is particularly suited for analytical workloads involving large volumes of text data, allowing users to efficiently perform searches and aggregations over billions or even trillions of rows, as exemplified by its application in platforms like Ryft.io and Icite. Despite the increased storage overhead of text indexes compared to Bloom filters, the trade-off is offset by substantial query performance improvements. ClickHouse's implementation allows customization through tokenization and pre-processing options, enabling users to tailor searches to specific requirements. The introduction of this feature aligns with ClickHouse's strengths as a high-performance analytical database, enhancing capabilities for analytics and observability workloads.

GitTrends is an open-source demo application designed to function like Google Trends for the tech industry, allowing users to search and compare trends across over 10 billion GitHub events in real time. Built to demonstrate the full-text search capabilities of ClickHouse, GitTrends enables users to track the popularity of technologies and topics by analyzing mentions in issues, pull requests, and comments on GitHub. It offers insights into tech adoption by allowing comparisons of keyword trends, identification of key repositories and ecosystems, and exploration of the specific conversations driving these trends. The application showcases the power of ClickHouse's new full-text search feature, providing instant search results by indexing raw text and enabling aggregation within the same engine. GitTrends, which is fully open-source and available for exploration, also includes features like live query performance comparisons and opportunities for users to inspect and run the SQL queries behind each chart.

The text discusses the challenges of maintaining consistent metrics across various platforms like chat tools, APIs, and dashboards, particularly when using a database like ClickHouse. It highlights the problems of inconsistent metric definitions and the need for a unified approach to ensure accuracy and reliability. The proposed solution involves implementing a lightweight metrics layer, or semantic layer, using an open-source tool called MooseStack. This approach allows metrics to be defined once in code and projected consistently across all platforms, ensuring that changes are synchronized and reducing the risk of errors. The text emphasizes the importance of treating the query model as a contract for production analytics, where metrics are reviewed, tested, and deployed like code to maintain consistency and reliability. It suggests that while ad hoc analysis has its place in development and exploration, production environments should rely on standardized, code-based metrics definitions to ensure accuracy and efficiency. The post also provides resources and guidance for implementing this approach, including a demo app and documentation to help users get started.

Replo utilizes ClickHouse to facilitate real-time, in-product analytics for Shopify merchants, enabling them to track live page interactions, offers, and A/B testing outcomes. Initially, Replo faced challenges with its analytics pipeline, such as inefficiencies in data recomputation and session-level metric redundancies, which led to the development of a more structured data model incorporating customer-specific namespaces and precomputed metrics. This approach reduced the computational load and improved query performance, allowing for the processing of over 100 billion events while maintaining responsive dashboards and accurate real-time analytics. The system evolved further to handle complex requirements like fractional attribution, eventually leading to a refined architecture that focuses on real-time session events to avoid the inefficiencies of full historical dataset scanning. As a result, Replo's analytics infrastructure became more scalable and reliable, supporting continued performance optimization and future enhancements.

When using LLMs (Large Language Models) to design ClickHouse tables for real-time event analytics, users may encounter several pitfalls if they rely solely on AI-generated schemas without human validation. The text highlights common mistakes such as inappropriate partitioning, overuse of custom codecs, unnecessary projections, and mismanaged JSON columns, which can lead to inefficiencies and performance issues at scale. It emphasizes the importance of starting with simple schemas, understanding the rationale behind AI-generated decisions, and adding complexity only when justified by actual workload measurements. The text also advises consulting human experts for complex scenarios and large-scale operations, noting that while LLMs are helpful for getting started, human insight is crucial for nuanced, high-stakes decisions. As AI tools continue to improve, collaboration between AI and human expertise will become increasingly important in database design and operation.

With the introduction of version 26.2, ClickHouse now embeds the ClickStack UI directly into its binary, enabling users to effortlessly explore observability on their local instances. Previously available primarily through Docker or as a managed service on ClickHouse Cloud, this integration allows users to navigate to a local URL to begin analyzing logs, traces, and metrics instantly. Designed for local exploration and learning, the embedded UI simplifies the setup process and reduces the need for additional services or dependencies, though it lacks features like persistent storage and alerting for production use. The integration achieves a minimal footprint by embedding static resources and using efficient gzip compression, sustaining ClickHouse's lightweight nature. Despite the absence of dynamic server-side rendering due to the constraints of embedding a full web application in a C++ binary, the static pages powered by Next.js support local data exploration effectively. This innovation aims to lower the entry barrier for users wanting to experiment with ClickStack and understand ClickHouse’s performance, though larger deployments are still recommended to use separate ClickStack components.

ClickStack's integration with ClickHouse Cloud API enables teams to streamline observability configurations across multiple services and environments by embedding them into deployment pipelines, version control, and infrastructure-as-code workflows, reducing manual work and the risk of inconsistencies. Users can now manage core ClickStack resources programmatically, including creating, reading, updating, and deleting dashboards and alerts via API, with dashboards retaining consistent layout and behavior across environments. The API also supports role-based access control and includes a clean, predictable specification for generating SDKs and integrating with CI/CD tools. Future developments include a Terraform provider, more granular access controls, and expanded API resources. Teams can start using ClickStack rapidly, provided they have the necessary ClickHouse Cloud organization and API key permissions, and are encouraged to provide feedback and suggestions for future features through dedicated channels like ClickHouse Slack.

Lago, a French startup, utilizes ClickHouse Cloud to efficiently manage large-scale, real-time usage-based billing for complex enterprises by scaling its event ingestion from 10,000 to 1 million events per second. Originally, Lago's founders discovered the intricacies of billing systems while working at fintech company Qonto, realizing the importance of transparent, open-source systems that avoid the limitations of proprietary billing solutions. ClickHouse Cloud was chosen for its ability to handle high ingestion rates, fast analytical queries, and activity logs without requiring extensive configuration or maintenance efforts. This infrastructure supports Lago's goal of providing flexible billing solutions for businesses of varying complexity, including those with AI-driven and API-first models, by allowing accurate, low-latency billing and real-time usage insights. As Lago expands its customer base to larger enterprises, ClickHouse continues to be integral, enabling scalable and reliable performance as Lago aims to consolidate its data management around a single high-performance system.

ClickHouse.Driver 1.0.0 is the first stable release of the official .NET client for ClickHouse, featuring a newly designed API, full type coverage, package signing, and OpenTelemetry integration. This release, which builds on previous versions 0.8.0 and 0.9.0, introduces the ClickHouseClient API, which replaces most use cases for ADO.NET classes, offering a more streamlined, thread-safe, and singleton-friendly experience. It includes over 40 practical usage examples and enhanced documentation. Key updates include automatic parameter type extraction from SQL, JWT authentication, roles, custom HTTP headers, and improved POCO serialization for JSON columns. Mid-stream exception detection and query ID auto-generation are also introduced, with notable breaking changes such as dropping support for .NET Framework/.NET Standard and modifying DateTime behavior for columns without explicit timezones. The release is focused on enhancing usability, performance, and compatibility, with migration guidance provided for adapting to the new version.