Datadog | The Monitor blog

Control trace volume with OpenTelemetry tail-based sampling

OpenTelemetry tail-based sampling reduces observability costs and trace volume by selectively retaining high-value data, such as errors and latency spikes, after a trace completes.…

From signals to systemic risk: Building Risk AI

Datadog has developed a Systemic Risk Detection Pipeline and Risk AI Agents to shift security focus from isolated findings to interconnected "risk paths" that pose the greatest org…

Detect vulnerabilities in LLM applications with Datadog’s AI-native SAST

Datadog Code Security’s AI-native SAST addresses the unique security risks of LLM applications, such as prompt injection and excessive agency, which traditional pattern-based tools…

20× the CI traffic without getting slower: How we rebuilt Git serving at Datadog

To handle the massive Git traffic and high CPU load caused by large monorepos and AI coding agents, Datadog developed "gitretriever," a decentralized system of independent Git mirr…

How CISA’s BOD 26-04 changes vulnerability prioritization

CISA’s BOD 26-04 mandates that federal agencies prioritize vulnerability remediation based on specific risk factors, such as asset exposure and exploitability, to counter increasin…

Trace AWS Lambda durable functions with Datadog

Datadog’s Node.js and Python tracers for the AWS Lambda Durable Execution SDK enable seamless, cross-invocation tracing for long-running, multi-step workflows. By stitching togethe…

Centralize human and agentic work with Datadog Work Management

Datadog Work Management provides a centralized platform to track operational tasks from humans, automations, and AI agents, addressing the fragmentation caused by using disparate t…

Two ways to measure the cumulative impact of experiments

Simply summing the observed lifts of winning experiments leads to an overestimation of their true impact, a phenomenon known as the "winner’s curse." To achieve more accurate resul…

Data pipeline monitoring 101: Tracking health and performance across the data stack

This article outlines the necessity of end-to-end monitoring to ensure data quality and integrity within the complex, multi-layered structures of modern data pipelines. It provides…

Avoid Azure secret rotation with secretless authentication

Datadog has introduced secretless authentication for its Azure integration, replacing manual client secret management with an identity-based model using federated credentials throu…

Investigate account-level churn risk with Product Analytics account segments

Datadog Product Analytics now supports account segments, allowing teams to combine business context—such as ARR and renewal dates—with user behavioral data. This feature shifts the…

How we improved APM Java startup by encoding a prefix trie as a JVM constant

To improve Java application startup performance, Datadog's APM team reduced class-matching overhead by 30% by optimizing how classes are selected for instrumentation. They achieved…