Cluster Resources

The cluster resources section shows how CPU, memory, and GPU capacity is allocated across all workloads.

CPU Allocation

The CPU pie chart shows:

• Used — sum of CPU requests from all running pods
• Free — remaining allocatable CPU (total allocatable minus requests)

CPU is measured in cores. Fractional cores (e.g., 0.5 = 500 millicores) are supported.

Memory Allocation

The memory pie chart shows:

• Used — sum of memory requests from all running pods
• Free — remaining allocatable memory

Memory is measured in GB.

GPU Resources

The GPU card shows total GPU units available across all nodes. Nodes with GPUs are detected automatically if they carry the nvidia.com/gpu resource label.

Metric	Description
Available	Total GPU units across all GPU nodes
In Use	GPU units currently allocated to pods

Resource Requests vs. Actual Usage

Note

The pie charts reflect Kubernetes resource requests (what pods asked for), not actual CPU/memory consumption. A pod might request 2 CPU but only use 0.3 CPU. For actual utilization, see Node Monitoring.

Resource requests are important for:

• Kubernetes scheduling decisions (will my pod fit on this node?)
• Capacity planning (am I over-provisioned?)

Actual usage is important for:

• Performance monitoring (is my workload CPU-bound?)
• Cost optimization (am I wasting resources?)