How does the Cluster Agent monitor Kubernetes environments?

The Cluster Agent runs as its own pod inside a Kubernetes cluster and talks to the Kubernetes API server to collect cluster-level state and events — pod status, deployments, node conditions, and resource utilization — giving visibility into the orchestration layer itself, on top of whatever the in-process App Agents inside individual pods already report.

This split matters because App Agents alone can tell you a service is slow, but not why a pod was suddenly evicted, hit a memory limit, or failed a readiness probe; the Cluster Agent supplies that orchestration-layer context, so a transaction slowdown can be correlated with a pod restart or a node running low on allocatable resources happening at the same moment.

Together, the Cluster Agent's infrastructure view and the App Agent's transaction-level view let a team distinguish "the code got slower" from "Kubernetes is throttling or evicting the pods running the code," which are very different problems requiring very different fixes.