ClusterHealth

Purpose

Provides a comprehensive overview of the entire Kubernetes fleet, tracking connectivity, health status (Ready/NotReady), authentication state (expired tokens), and resource capacity (CPU/Nodes/GPUs).

Data Sources

• Primary Hook: useClusters() (Retrieves aggregated cluster telemetry)
• Support Hooks:
  • useCachedGPUNodes: Calculates per-cluster and global accelerator allocation.
  • useGlobalFilters: Scopes the "Summary Stats" to the user's selected clusters.
  • useCardData: Standard filtering (search name, context, server), sorting, and pagination.
  • useCardLoadingState: Coordinates skeletons until health checks complete.
• Provider Priority: (Backend REST → Backend SSE → Local Agent Proxy → Demo Data)

Props / Configuration

Prop	Type	Default	Description
config	object	undefined	Optional scoping (e.g., { cluster: string, namespace: string }).

UI States

• Loading: Custom skeleton with header, stats grid, and list shimmer.
• Empty: noClustersConfigured message with onboarding prompt.
• Error: Inline warning box for connectivity or re-authentication (expired tokens).

Logic & Patterns

• Health Determination: Healthy if nodeCount > 0 OR healthy === true.
• Status Mapping: Clusters are categorized as "Healthy", "Unhealthy", "Auth Error" (key icon), or "Offline" (wifi icon).
• Sorting: Supported by status, name, nodes, and pods.
• AI Integration: CardAIActions triggers "AI Diagnose & Repair" for any unreachable or unhealthy clusters.
• Drilldown: Clicking a cluster name opens the detail modal for deep-dive metrics.

Notes

• Performance: Performs parallel health checks across all context names (aliased vs. raw paths).
• Edge Cases: Handles cloud provider detection (EKS, GKE, AKS, OpenShift) via server URL analysis to provide direct links to the cloud console.