Best Practices for Deploying AWS DevOps Agent in Production

This article provides best practices for deploying AWS DevOps Agent in production, focusing on configuring Agent Spaces to balance investigation capability with operational efficiency.

• Agent Spaces define resource access boundaries for autonomous incident investigation
• Design Agent Space boundaries aligned with on-call team responsibilities and application architecture
• Separate production from non-production environments in different Agent Spaces
• Support three patterns: single applications, shared services teams, and central operations at scale
• Use Infrastructure as Code (CDK/Terraform) for consistent, repeatable deployments
• Configure IAM roles across accounts to enable cross-account resource access
• Integrate observability tools: CloudWatch, APM tools, code repositories, CI/CD pipelines
• Support webhooks and MCP servers for custom monitoring tool integration
• Implement fine-grained IAM-based access controls for investigations
• Test Agent Space configuration and iterate based on investigation results

AWS DevOps Agent reduces incident resolution time by automating root cause analysis. Proper Agent Space configuration ensures investigations have sufficient context while maintaining operational efficiency and security boundaries.