Summary Multi-Agent Reference Architecture Introduction Building blocks Multi-agents design options Agents as a modular monolith Agents as microservices Agents registry Agents communication Request-based Message-driven Data exchange protocols Memory Memory architecture patterns Short-term memory Long-term memory Memory pipelines Observability Evaluation Tool Call Evaluation Security Governance Experimentation to productization DevOps and DataOps on agents lifecycle Context Engineering Iteractive Optimization Loop Agents Orchestration Tools Context for Agents Reference architecture Patterns Contributors References