January 2026 Summaries

Datadog's Cloud Cost Management (CCM) team developed a natural language query (NLQ) agent that translates plain-English questions into valid Datadog metrics queries, allowing FinOps and engineering users to evaluate costs with ease. The agent's non-conversational nature required a focus on correctness, prompting the team to conduct user testing and create a reference dataset from real user prompts. To address the challenges posed by the nondeterministic nature of large language models, Datadog implemented LLM Observability with component-level evaluators for parsing, metric selection, roll-up, group-bys, and filters, enabling more precise debugging and iteration. This approach streamlined testing and debugging, reducing time spent on these tasks by 20 times through automated evaluations and trace-level inspection. Additionally, the use of Datadog's distributed tracing facilitated seamless integration with existing systems, allowing for objective model comparisons and continuous improvement of the NLQ agent.

Datadog's Database Monitoring now supports the automatic collection of EXPLAIN ANALYZE plans for PostgreSQL, facilitating the troubleshooting of slow queries by providing detailed execution insights and correlating them with APM traces. This feature allows users to detect performance issues, identify root causes, and diagnose specific problems directly within the Datadog platform. It provides an interactive visualization of execution plans, highlighting critical nodes for easier interpretation, especially for users unfamiliar with explain plans or large queries. The tool can help identify discrepancies in estimated versus actual rows, which may indicate stale table statistics necessitating an ANALYZE operation, or reveal if high disk reads are causing latency, suggesting the need for improved caching or additional RAM. By offering these capabilities, Datadog enables a seamless transition from detecting application performance issues to resolving query latency problems, all within a single interface.

Generative AI has accelerated the pace of software development, but traditional testing methods have struggled to keep up, particularly in evaluating AI outputs. To address this challenge, Propolis, an autonomous QA testing platform, has been developed to focus on goal- and output-oriented testing by using synthetic users to explore applications and identify real user journeys and goals. Propolis autonomously updates and heals tests, ensuring comprehensive coverage without human intervention. Now joining Datadog, Propolis will integrate with Datadog's production context tools like traces, logs, and Real User Monitoring data to fully automate end-to-end testing. This integration aims to shift quality assurance from a pre-release checkpoint to a continuous signal within CI/CD and runtime, providing a more dynamic and responsive approach to software testing.

Datadog has introduced new RUM x APM cross-product retention filters, available in Preview, to enhance the correlation between frontend and backend data for improved application performance troubleshooting. These filters enable teams to retain all backend Application Performance Monitoring (APM) traces associated with retained Real User Monitoring (RUM) sessions, thereby providing a unified view of the user journey from browser events to backend service calls. By setting up cross-product retention filters, teams can easily diagnose issues like application errors or slowness that often stem from backend problems, such as database bottlenecks. The feature allows for efficient investigation of critical user experience moments by retaining sessions with relevant traces, helping identify the services or dependencies contributing to failures without additional configuration. Datadog offers a default 1% cross-product retention filter to ensure a representative sample of correlated sessions and traces, allowing users to analyze cross-stack issues at no extra cost. This integration of RUM and APM data facilitates faster troubleshooting and improved productivity by connecting frontend and backend observability, ensuring a consistent end-to-end view of factors affecting user experience.

As cloud infrastructures expand, maintaining compliance with regulatory and industry standards becomes increasingly challenging due to rapid changes in cloud resources, decentralized service adoption, and dispersed audit evidence. Datadog Cloud Security addresses these challenges by providing a unified, continuous approach to compliance, offering extensive framework coverage, real-time visibility into posture risks, and over 1,500 out-of-the-box detection rules tailored to industry standards. It supports both agentless and agent-based scanning, ensuring near real-time insights and a 15-month history of misconfigurations to aid audits and investigations. Users can customize controls using Rego, with each misconfiguration tagged for compliance relevance, thereby simplifying the understanding and prioritization of risks. Datadog also supports various global and industry-specific frameworks, including APRA CPS 234, SOC 2, PCI DSS 4.0, and ISO 27001:2022, enabling organizations to maintain audit readiness and meet diverse compliance requirements. The platform facilitates continuous monitoring and control mapping, enhancing security posture by minimizing manual compliance processes and integrating orchestration tools for faster issue resolution.

As organizations expand their cloud environments and branch office networks, maintaining reliable connectivity becomes increasingly complex, which VeloCloud SD-WAN addresses through dynamic, policy-based routing to ensure dependable and cost-efficient connectivity and consistent application performance. Datadog's integration with VeloCloud SD-WAN provides unified visibility into network connectivity by collecting metrics and logs directly from the VeloCloud Orchestrator, allowing for proactive WAN management and faster issue detection. This integration supports monitoring of link performance and edge device resource usage, enabling diagnostics of network disruptions through real-time anomaly detection and customizable dashboards. Additionally, Datadog Network Device Monitoring (NDM) offers a device-oriented view, facilitating the exploration and investigation of network hardware performance, while machine learning-powered anomaly detection identifies underperforming transport links to optimize routing policies. The combined capabilities of NDM and the VeloCloud integration deliver end-to-end visibility into SD-WAN infrastructure, reducing downtime and enhancing WAN reliability across distributed environments.

Datadog Observability Pipelines offers managed security service providers (MSSPs) a centralized and lightweight solution for ingesting and processing logs across diverse customer environments, addressing challenges at petabyte scale without the need for installed agents. It enables MSSPs to collect and standardize logs, apply enrichment, and ensure consistent security workflows, thereby reducing operational overhead and observability blind spots. The solution simplifies the migration process from legacy SIEM systems to modern platforms by supporting dual-shipping and phased transitions, allowing MSSPs to manage large-scale SIEM modernization projects more effectively. This approach allows MSSPs to enhance their log collection, processing, and routing capabilities while maintaining control and scalability, ensuring they can meet complex customer-specific taxonomy and compliance requirements.

Datadog's Graphing AI team has integrated LLM Observability into their widget and dashboard generation agents to enhance the reliability of AI-driven systems by providing real-time visibility into their behavior. This integration allows the team to trace agent interactions, debug complex chains, and evaluate semantic and functional accuracy across different model versions. By utilizing Datadog's Experiments feature, the team automates evaluation and structured testing at scale, enabling them to identify issues swiftly and maintain high performance. The agents, which convert natural language prompts into Datadog visualizations, now benefit from improved debugging tools, allowing engineers to pinpoint failures, such as HTTP 401 responses from service calls, and address them efficiently. This system enables a reproducible offline evaluation pipeline that measures accuracy through both deterministic checks and LLM-as-a-judge assessments. Future plans include expanding the framework to evaluate models from other providers like Anthropic Claude and correlating online evaluations with Real User Monitoring data to further refine user experience metrics.

On January 13, 2026, Node.js disclosed a denial-of-service (DoS) vulnerability, identified as CVE-2025-59466, affecting applications using async_hooks or AsyncLocalStorage, potentially causing unexpected process termination in Node.js versions 8.x to 23.x. This vulnerability is significant for Datadog APM customers since the Node.js dd-trace-js tracer utilizes AsyncLocalStorage; however, the issue is not exclusive to Datadog's tracer, and validation against patched Node.js versions is ongoing. The vulnerability arises when a stack overflow occurs in user code with active async_hooks, bypassing standard error handlers and leading to unrecoverable process crashes, posing a risk to applications relying on request context, including those using React Server Components and Next.js. Remediation involves upgrading to patched Node.js versions released on the same day, with specific guidance available from Datadog's Security Research Feed, and Datadog is committed to ongoing monitoring and support for customers seeking configuration assistance.

Flaky tests present a significant challenge for engineering teams by disrupting development workflows, creating noise in CI/CD processes, and eroding trust in test suites. Datadog's Test Optimization tool addresses this by identifying flaky tests and analyzing their history to surface root causes, but fixing these issues has traditionally required significant developer effort. The integration of the Bits AI Dev Agent with Datadog Test Optimization enhances this process by autonomously generating verified code fixes, packaged as production-ready pull requests, upon detecting flaky tests. This AI-driven solution leverages rich historical data and execution traces to accurately diagnose problems, allowing developers to receive immediate, verified pull requests without having to manually address the issue. This integration aims to reduce the repetitive work of fixing flaky tests and foster developer trust by ensuring that the AI-generated fixes are pre-verified before being merged. By automating the resolution of test instability, engineering teams can reclaim time for more productive tasks, ultimately leading to more efficient development processes.

Bits AI SRE, developed to assist engineers in resolving production incidents in complex distributed systems, offers significant improvements in incident response times by autonomously analyzing telemetry data and providing root cause analyses. It mimics human Site Reliability Engineers by forming and testing hypotheses, focusing on causal relationships, and conducting deep investigations to identify the root causes of multi-component issues. By evaluating its performance against real-world incidents using the extensive telemetry dataset from Datadog, Bits AI SRE has shown marked improvements, with the capability to significantly reduce noise and focus on relevant data. The tool continues to evolve, integrating with more expert investigation and optimization agents within the Datadog platform, allowing it to cover a broader range of real-world scenarios and drive comprehensive resolution workflows. Users have reported positive feedback, noting a reduction in the time required to detect root causes, and the tool is continually expanding its capabilities.

In 2025, Datadog marked a significant milestone by surpassing 1,000 integrations, expanding its ecosystem with over 110 new technology partners and enhancing coverage across emerging software categories like AI, distributed security, hybrid infrastructure, and data intelligence. The year witnessed a focus on AI observability and cost control, with integrations such as GitHub Copilot and Microsoft Copilot providing insights into AI lifecycle management and cost tracking. In response to evolving security challenges from increased SaaS adoption and cloud services, Datadog expanded its Cloud SIEM capabilities with integrations that offer comprehensive visibility into cloud environment risks and threat actor activities. To support the growing adoption of hybrid architectures, Datadog enhanced monitoring for distributed systems, covering on-prem, cloud, and edge environments with new integrations like Proxmox and Temporal Cloud. Additionally, Datadog extended its visibility into data and analytics systems, bolstering business insight capabilities with integrations for tools like dbt Cloud and Tableau, while also improving monitoring for digital experience platforms like Shopify and Mailchimp.

In an organization, leaders often face challenges in synthesizing data from disparate sources to answer critical business questions effectively. Creating executive dashboards in Datadog can address this by integrating various telemetry data, including product usage and customer experience metrics, to provide a cohesive narrative that enables quick, informed decision-making. The process involves understanding the audience, preparing data with consistent tagging and naming conventions, and applying simplicity-first design principles to ensure clarity and relevance. High-altitude dashboards give executives a concise overview of key metrics like revenue and market trends, while lower-altitude dashboards offer detailed insights into operational performance and user behavior. These dashboards should be easy to navigate, structured logically, and adaptable through template variables to meet different stakeholders' needs. An example of an ecommerce executive dashboard illustrates how connecting customer behavior, revenue, and system reliability in a single view can help prioritize actions based on business impact. By doing so, Datadog helps organizations transform raw data into strategic insights, aligning operational health with overall business outcomes.

Comprehensive observability is essential for maintaining efficient and secure serverless workloads, but scaling and sustaining visibility across numerous serverless functions can be challenging, particularly when teams use various infrastructure as code (IaC) frameworks and deployment processes. Datadog addresses these challenges by offering a centralized interface for managing AWS Lambda functions, enabling consistent monitoring without requiring code updates. By deploying a remote instrumenter CloudFormation stack, developers can automate the instrumentation process within minutes, ensuring telemetry flows seamlessly and functions are continuously monitored. This approach allows for easy selection and configuration of multiple Lambda functions using tags, maintaining up-to-date settings for tracing and logging, and automatically instrumenting new functions. Datadog's remote instrumentation for AWS Lambda functions offers a streamlined way to balance speed and visibility, ensuring data is available for troubleshooting and monitoring across different teams and services.

eBPF has transformed the landscape of observability, networking, and security by enabling deep kernel-level instrumentation. However, deploying eBPF at scale, especially in diverse production environments, uncovers numerous challenges and complexities. Over five years, Datadog's experience with its Workload Protection product, which leverages eBPF, has revealed critical lessons for ensuring reliable operation. These include navigating edge cases across various kernel versions, ensuring comprehensive syscall coverage, managing performance impact, and maintaining secure and consistent data capture. Additionally, the potential for eBPF to be misused as a rootkit highlights the importance of strict monitoring and auditing. Datadog has developed strategies such as kernel version testing, centralized eBPF logic, and dynamic in-kernel filtering to mitigate these challenges. Despite its limitations, eBPF continues to hold significant promise for the future of systems engineering, though its adoption in managed or serverless environments remains uncertain. Datadog acknowledges the vast potential of eBPF and remains committed to leveraging its capabilities for various use cases while continuously refining its approach to maintain security and performance at scale.

dbt-expectations is an open-source package developed by Datadog that enhances dbt's core testing capabilities by introducing more sophisticated data quality checks to address common issues such as complex data validation, time series data quality, statistical validation, and cross-column validation. It is widely adopted by organizations to ensure data integrity and scalability by enabling Great Expectations-style assertions that go beyond dbt's native tests. The package allows users to implement comprehensive checks directly within their dbt models, providing an early warning system to prevent data processing errors that could compromise analysis. dbt-expectations supports integration with CI/CD pipelines, allowing continuous monitoring and testing to catch regressions before they cause issues in data quality. By focusing on business-critical data points and mapping tests to clear actions, users can effectively manage alert fatigue and ensure that alerts are actionable. The package is designed to work seamlessly within existing workflows, empowering data teams to maintain healthy and reliable data pipelines.

The GitLab Source Code integration with Datadog offers developers and SREs a streamlined approach to managing code quality and security by integrating GitLab repositories directly into Datadog's observability tools. This integration reduces the need for switching between different platforms by providing immediate visibility into code health, security vulnerabilities, and infrastructure misconfigurations directly within the development workflow. By embedding GitLab code snippets within Datadog's APM tools, it accelerates troubleshooting by allowing users to trace errors back to specific files without manual searches. Additionally, the integration automates code reviews by providing actionable feedback within GitLab merge requests, thus encouraging early detection of issues like flaky tests and regressions. This collaboration aims to enhance the efficiency of resolving code problems by suggesting fixes directly in the merge request interface, minimizing the context switching required for remediation. Overall, the integration fosters a unified, code-aware environment that helps teams prioritize and address issues swiftly before they affect end-users.