Dynatrace and Google Cloud have expanded their collaboration to provide observability capabilities for agentic AI workloads through two primary integrations: a Gemini CLI extension for developer access to observability data within terminal environments, and an A2A protocol integration with Gemini Enterprise for real-time system monitoring.

The integrations use MCP servers to expose Dynatrace telemetry data to AI agents, creating a feedback mechanism where observability platforms function both as monitoring tools and validation sources for autonomous agent behavior.

Technical Details

Dynatrace has implemented two distinct integration paths for Google Cloud’s agentic AI platforms, each targeting different operational contexts within enterprise environments. T

he technical foundation relies on MCP servers that function as intermediaries between Dynatrace’s observability platform and Google’s AI agent frameworks.

The core technical components include:

• Gemini CLI Extension: Provides terminal-based access to Dynatrace observability data and root-cause analysis capabilities. Developers working within command-line environments can query application performance metrics, trace distributed transactions, and access diagnostic information without switching to separate monitoring interfaces.
• Gemini Enterprise A2A Integration: Enables direct communication between AI agents and the Dynatrace observability platform using Google’s Agent-to-Agent protocol. This allows AI agents to programmatically access telemetry data, submit queries about system state, and receive contextual information about application performance. [The A2A protocol, now managed under the governance of the Linux Foundation, establishes standardized communication patterns between autonomous systems.]
• MCP Implementation: Dynatrace operates MCP servers that expose telemetry data to AI agents in structured formats. [MCP, governed by the Agentic AI Foundation under the Linux Foundation umbrella, defines how context information flows between AI systems and data sources.]

Observability Capabilities for Agentic Systems

Traditional observability approaches designed for human-initiated workflows do not provide adequate visibility into agent decision-making processes, task handoffs between agents, or the cascading effects of agent actions across system boundaries.

Dynatrace’s integrations address these technical challenges, enabling observability for AI agents operating autonomously across extended timeframes and complex distributed environments.

Dynatrace provides:

• Full-Stack Visibility Across AI Workflows: Telemetry collection spans application code, infrastructure resources, network transactions, and external service dependencies. This allows DevOps teams to trace the complete execution path of agent-initiated actions, from initial trigger through final completion.
• Root Cause Analysis for Agent Disruptions: The platform applies automated analysis algorithms to identify causal relationships between events that precede system disruptions. For agentic AI workloads, this capability enables teams to determine which agent actions, resource constraints, or external dependencies caused workflow failures.
• Real-Time Telemetry Access for Adaptive Agents: AI agents can query current system state and performance metrics to inform decision-making processes. An agent orchestrating container deployments, for example, could access resource utilization data before scaling decisions, creating feedback loops where agents adjust behavior based on observed system conditions (rather than operating from static configurations).

Analysis

The collaboration between Dynatrace and Google Cloud on agentic AI observability addresses real technical requirements emerging as enterprises deploy autonomous agents across production environments. Long-running AI agents operating without continuous human oversight require instrumentation that traditional monitoring approaches do not provide.

Observability for agentic AI extends beyond monitoring agent health. Observability solutions must evolve to validate agent decisions, audit agent actions, and provide the contextual telemetry data agents require for adaptive behavior.

This expanded scope transforms observability into an architectural foundation that enables safe, effective agent autonomy. The question facing IT leaders is whether their current observability solutions can scale to meet the dramatically increased demands these architectures create.

As enterprises increasingly delegate operational tasks to AI agents, observability platforms become the essential control plane that enables organizations to trust, govern, and optimize autonomous systems operating at machine speed across complex distributed environments.

Dynatrace, with its new Google Cloud integrations, provides a strong path forward for enterprises committed to Google’s AI platforms.

Competitive Outlook & Advice to IT Buyers

Dynatrace competes against multiple observability platforms pursuing similar agentic AI integration strategies, including Datadog, New Relic, Cisco’s Splunk, and a few emerging AI-native observability platforms.

Let’s look at how this collaboration stacks up…

These sections are only available to NAND Research clients and IT Advisory Members. Please reach out to [email protected] to learn more.