AWS re:Invent 2025 - Move beyond reactive: Transform cloud ops with AWS DevOps Agent (COP362)

AWS DevOps Agent: Transforming Cloud Operations

Introduction to the AWS DevOps Agent

• The AWS DevOps Agent is a new service launched by AWS at re:Invent 2025 to help teams improve cloud operations and incident response.
• It is designed to work autonomously as part of the on-call team, responding to incidents, analyzing root causes, and providing mitigation steps.
• The agent is built on the principles of DevOps and observability, leveraging a wide range of telemetry sources and integrations.

Key Capabilities of the AWS DevOps Agent

Incident Response and Resolution

• The agent can rapidly analyze metrics, logs, traces, and other telemetry to identify the root cause of an incident.
• It generates a detailed incident report with the root cause analysis and recommended mitigation steps.
• The agent can even provide the specific code changes or commands needed to resolve the issue.

Proactive Incident Prevention

• The agent continuously scans past incidents to identify patterns and opportunities for improvement.
• It provides recommendations to fix underlying issues and enhance the overall operational posture of the application.
• Recommendations cover areas like infrastructure, observability, governance, and deployment pipelines.

Flexible Telemetry Integration

• The agent integrates with a wide range of observability tools, including AWS services, commercial platforms, and open-source solutions.
• It can also connect to custom "bring your own" telemetry sources through an MCP (Managed Control Plane) server integration.
• This flexibility allows the agent to work seamlessly with an organization's existing observability stack.

Collaborative Incident Management

• The agent is designed to work as a member of the on-call team, communicating findings and recommendations through channels like Slack, ServiceNow, and more.
• It can also bring in human experts, like AWS support engineers, to collaborate on complex incident investigations.
• Users can interact with the agent through a chat interface, asking questions, steering investigations, and accessing detailed reports.

Technical Implementation and Architecture

• The agent builds an application topology by discovering resources, their relationships, and relevant telemetry sources.
• It uses this topology, along with integrations to CI/CD pipelines, to rapidly analyze incidents and identify root causes.
• The agent's capabilities are further enhanced through the use of "steering files" or runbooks, which provide organization-specific guidance and best practices.
• The agent operates as a secure, isolated service within the customer's environment, with access controlled through IAM roles.

Customer Adoption and Results

• The AWS DevOps Agent was initially tested internally within Amazon and AWS, handling over 1,000 incidents with an 86% success rate on root cause analysis.
• Early beta customers, like the Commonwealth Bank of Australia and Western Governors University, have seen significant improvements in incident response times and the ability to prevent future issues.
• The agent's flexibility and integrations with tools like Dynatrace have enabled joint customers to quickly set up and deploy the solution, realizing benefits within a single day.

Conclusion and Next Steps

• The AWS DevOps Agent represents a significant advancement in cloud operations, leveraging the latest in observability, automation, and artificial intelligence.
• By automating incident response and providing proactive recommendations, the agent can help organizations transform their cloud operations, reducing mean time to resolution (MTTR) and preventing future incidents.
• The public preview of the AWS DevOps Agent is now available, and customers are encouraged to try it out and provide feedback to help shape the future development of the service.