Building a Kubernetes Graph Engine for Agents

Kubernetes debugging often feels more like a procedural "treasure hunt" than a streamlined workflow. One simple question, like "what breaks if this node goes down?", usually turns into a dozen kubectl lookups across pods, services, and ingresses. Ariadne changes the game by turning live cluster state into a queryable Memgraph property graph.

In this technical session, Artavazd (Art) Balaian, Senior Lead Software Engineer at Agoda, reveals how he built Ariadne to replace procedural debugging with graph-based reasoning. You’ll see how Ariadne maintains a live Kubernetes state using incremental syncs and exposes that power to both engineers using Cypher queries and AI agents via the Model Context Protocol (MCP).