Node Monitoring

Node Monitoring

Node Monitoring shows infrastructure-level details for a node group (auto-scaling group or managed node pool). The page includes cost and efficiency metrics, analytical charts, and a list of individual nodes.

Access: Select cluster > Dashboard > Nodes tab > Click a node group

---

Four cards at the top:

• Nodes - Count of nodes in the group with trend indicator
• Monthly Cost - Combined monthly cost with trend indicator
• Average Efficiency - Average of CPU and memory efficiency across the group
• Total Resources - Aggregate CPU cores and memory

---

Four analytical charts provide insight into the node group:

• Pod Density Over Time - Bar chart showing how many pods are scheduled per node over time. Includes a formula explanation for density calculation.
• Cost per Node Type - Horizontal bar chart comparing costs across different instance types in the group.
• Availability Zone Distribution - Dual-axis bar chart showing node count and cost per AZ. Helps identify imbalanced distributions.
• Ghost Town Analysis - Area chart comparing provisioned capacity against actual utilization. Identifies off-hours waste where resources sit idle.

---

Two tabs organize node group data:

1. Nodes - Table of individual nodes. Columns: name, instance type, status, CPU/memory utilization, hourly cost, pods scheduled. Includes filters for search, instance type, spot/on-demand toggle, and GPU-enabled toggle.
2. Details - Infrastructure metadata including provider, instance type, region, zones, architecture, OS, GPU availability, and total CPU/memory/pod capacity.

---

Each node in the table shows:

• Name - Node hostname
• Instance Type - EC2/GCE instance size (e.g. m5.2xlarge, t3.large)
• Status - Ready, NotReady, or Unknown
• CPU Usage - Current utilization percentage
• Memory Usage - Current utilization percentage
• Hourly Cost - Total hourly cost for this node
• Pods - Number of pods scheduled on this node

For efficiency score calculations, see Resource Efficiency.