On the first episode of This Month in Datadog in 2026, Jeremy covers how you can protect agentic AI applications with AI Guard, stay up to date and collaborate during incidents with five Incident Management releases, and ship software with confidence using Feature Flags.
Later in the episode, Kevin spotlights Datadog Data Observability, which enables you to detect data quality and pipeline issues early. Also in this episode: Learn about the Datadog Ambassador program and upcoming Datadog Summits in London on March 10 and Sydney on March 17.
New features
Catch silent data issues early with Data Observability
Teams need data they can trust when building AI and analytics. Datadog Data Observability enables you to detect data and pipeline issues early and remediate those issues with end-to-end data lineage. You can also see an issue’s scope of impact to identify which AI models or BI dashboards are at risk, as well as diagnose the upstream root cause.
Protect agentic AI applications from real-time threats with AI Guard
The abilities of today’s agentic AI applications are powerful, while also introducing new opportunities for misuse, data exposure, and prompt-driven manipulation. Datadog AI Guard can evaluate prompts, responses, and tool calls, and block requests that appear harmful before they reach critical systems and sensitive data.
Stay up to date and reduce steps to collaborate during incidents with new releases
We’re helping you respond to incidents with five new Incident Management releases. Create AI-powered incident summaries with one click in Datadog or by running /datadog incident summary in Slack. Automatically create Google Chat spaces or Microsoft Teams meetings when incidents are declared. Find and filter incidents more easily with wildcard searching, enhanced query grammar, and expanded free-text search.
Release software with confidence using Datadog Feature Flags
Now generally available, Datadog Feature Flags lets you create and manage features where you already monitor application health. Create flags and variants with values of any data type, define targeting rules for rollouts, and configure flags to roll out instantly or gradually. Because flags are tied to telemetry data, Datadog can automatically pause or roll back a release if performance degrades or errors spike.
Additional updates
More new features and updates released this month:
- Unify frontend and backend data with cross-product retention filters
- Detect human names in logs using Sensitive Data Scanner
- Identify untested code early and more with Code Coverage
- Automatically upgrade Datadog Agents using Fleet Automation
See you next month
This Month in Datadog is a monthly roundup of our latest features, product announcements, and more. Subscribe to our YouTube channel to get notified when future episodes are live.
In the meantime, check out our release notes for a full list of new features and updates. Or see them in action by logging in to the Datadog platform or signing up for a 14-day free trial. See you next month!
