Detect when your AI agent copies a self-replicating prompt injection
Datadog AI Guard can now flag self-replicating prompt injections where an AI agent copies hidden instructions into files or email replies.
Datadog AI Guard can now flag self-replicating prompt injections where an AI agent copies hidden instructions into files or email replies.
Datadog introduced the Infrastructure Basic tier, a lightweight monitoring option designed for lower-priority workloads to track host health and resource usage.
Datadog introduced Agent Observability to help engineers identify applications driving AI spend, prioritize savings, and validate cost changes against performance.
Datadog launched Scout 1.0, an AI model specialized for investigating production issues. Powering Bits AI, Scout 1.0 provides improved price-performance compared to general-purpose frontier models on target SRE scenarios.
Datadog released cross-cluster queries for Datadog BYOC Logs, enabling AI agents and engineering teams to search logs across regions and business units without centralizing log storage.
Datadog Incident Response is now available in FedRAMP High environments, allowing federal agencies to handle paging, incident coordination, and automated response workflows for sensitive workloads.
Datadog Code Security now supports Supply Chain Firewall, providing centralized management of package allowlists and blocklists. It also includes retroactive scanning of installed packages to block malicious dependencies.