DDOG - DASH 2026 Recap: Observing a Normal

Research evidence excerpt

DDOG - DASH 2026 Recap: Observing a Normal

privileges, access sensitive data, and

communicate externally, and a single malicious instruction hidden inside an innocuous-

looking prompt can turn a well-intentioned agent into a malicious actor. AI Guard combines

deep agent telemetry tracing with stateful AI-native behavioral anomaly analysis to catch

multi-step, agent-poisoning attacks that stateless evaluators miss. The family extends

across the agent lifecycle, with AI Guard for Coding Agents protecting the dev-time surface

(Cursor, Claude Code, Copilot), and AI Guard Agent Discovery auto-mapping every agent

operating in the environment. Datadog also launched infinite-cardinality metrics, purpose-

built for agentic querying so customers can capture every attribute and dimension, priced

per metric and scaling with the data rather than penalizing depth.

Consolidating Network Monitoring. Datadog Network Monitoring went GA at DASH 2026,

paired with Network Device Monitoring to provide visibility into both the traffic flowing through

the network and the underlying devices powering it. A new current vs. prior traffic route

comparison view makes path changes and regressions visually obvious, allowing teams to

quickly spot when traffic is moving differently than it should. The workflow ties directly back

into Bits AI, so a network anomaly can be investigated, correlated to application impact,

and routed toward remediation inside the same platform, extending the "one platform, one

investigation" pitch to the network layer. To our understanding, the Network Monitoring

announcement bundles the end-to-end Network Monitoring suite that correlates network,

application, and infrastructure telemetry in real time across both cloud and on-premises

environments.

Positive Customer Checks.