Operations | Monitoring | ITSM | DevOps | Cloud

Improving an AI agent means knowing which behaviors to change and where those changes will have the greatest impact. Insights in Datadog Agent Observability improves AI agents by analyzing agent traces to identify recurring cost and reliability problems. It then explains the causes of those problems, assesses their impact, and recommends fixes backed by evidence. At Datadog, we use Insights to improve the agent behind Insights.

In September 2026, OpenAI reported finding self-replicating prompt injections. These attacks hide instructions in content an AI agent retrieves, directing it to take unauthorized actions and copy those instructions into emails, files, or messages. Copied instructions can then expose other agents or later sessions to the same attack, giving it the potential to spread like a computer worm.

Datadog has been named a Leader in The Forrester Wave: AIOps Platforms, Q4 2026. We believe this placement reflects Datadog’s commitment to transforming observability into autonomous operations with a single integrated platform where AI agents detect, investigate, and remediate production issues, while engineers stay in control. Close dialog.

A voice agent can return the right words and still deliver a poor experience. It might leave a caller in silence, speak over them, or stop midway through a response. A transcript may look correct even when the audio the user heard was delayed, incomplete, or different from the text on screen. These problems are difficult to diagnose from text traces alone. Voice conversations also involve more than a single request and response.

Serverless should be the first execution model teams evaluate for AI agent workloads. Many agents spend much of their lifetime waiting, and always-on compute forces teams to pay for every idle minute. In a study of more than 35,000 production coding-agent sessions, the median session ran LLM or tool calls for only 20% of its lifetime. The rest of the time, it sat idle between steps, with its sandbox still allocated.

A back-office server that runs out of disk space can interrupt an important internal workflow. A process that stops on a staging host can delay development. Although workloads like these might not require application-level instrumentation, teams still need to know when the hosts that support them require attention. Without a monitoring solution tailored to these requirements, teams have to apply more monitoring than the workload needs, maintain a separate tool, or accept a gap in visibility.

Running one A/B test well is manageable. Scaling A/B testing to hundreds of experiments across dozens of teams is harder, because teams may make different decisions about metrics, randomization, rollout, and analysis. Experiment protocols in Datadog Experiments turn an approved methodology into a reusable configuration, giving experiments of the same type a consistent starting point. In this post, you’ll learn how to.

Organizations can use Oracle AI Database@AWS, Oracle AI Database@Azure, and Oracle AI Database@Google Cloud to run Oracle databases co-located with the cloud applications they support rather than managing the databases on premises. However, running applications in a cloud provider’s environment while hosting databases on Oracle Cloud Infrastructure (OCI) can lead to visibility gaps. When application performance degrades, identifying the source of the problem can be difficult.

AI agent costs accumulate along more dimensions than a model-call count. Spend is a product of both which models you call and how the agent is architected: how many turns the loop runs, how many tools it invokes, how much retrieved context it carries, and how much of that context is re-sent on every turn. Provider bills tell you how much you spent, but they’re keyed to API keys and models instead of the applications and workflow steps that generated the calls.

Well-designed service level objectives (SLOs) help you track the benchmarks your app needs to meet to deliver a successful user experience (UX). To track performance against these SLOs, many SRE teams are turning to real user monitoring (RUM) and synthetic monitoring for their service level indicators (SLIs), as these forms of data provide more direct measurements of UX than backend performance metrics.