Evolving Policy Management with Agentic AI: Kyverno MCP and Kagent for... Shuting Zhao & Dahu Kuang

This talk covers evolving policy management through the integration of Agentic AI with Kyverno and Kagent for effective multi-cluster governance. The speakers, Shuting Zhao from Nirmata and Dahu Kuang from Alibaba Cloud, address the limitations of traditional policy management in dynamic environments and demonstrate how to leverage open-source tools to implement policy-as-code. They present a case study from Alibaba Cloud, highlighting the evaluation process for policy-as-code solutions and the extension of Kyverno’s capabilities with MCP servers for centralized management. The discussion emphasizes maintaining consistent governance across hybrid Kubernetes environments while tackling the complexities of large-scale policy frameworks.