kubeledger

Overview

KubeLedger tracks CPU, memory, and GPU usage across Kubernetes clusters over hourly, daily, and monthly intervals to provide precise per-namespace accounting and identify system overhead. With the inclusion of an optional MCP server, engineers and financial teams can use natural language prompts via supported AI clients to execute custom analytics, rank resource consumers, evaluate efficiency ratios, and review multi-period trend comparisons beyond rigid dashboard limits.

Capabilities

• ▸Track CPU, memory, and GPU resource usage and requests per namespace
• ▸Provide a read-only Model Context Protocol (MCP) server with ten specialized data tools
• ▸Analyze resource efficiency ratios and identify over or under-provisioned workloads
• ▸Generate hourly, daily, and monthly historical trend aggregations
• ▸Expose Prometheus compatible metrics for monitoring and alerting

Best for

Querying namespace resource consumption and top consumers using natural language prompts in an AI assistant, Conducting ad-hoc efficiency and capacity planning investigations for CPU, memory, and GPU utilization, Analyzing non-allocatable cluster overhead versus actual application workload capacity, Performing multi-period usage trend comparisons for cost allocation and chargeback reporting

Works with

Security considerations

The MCP server operates in a read-only mode without Kubernetes API write access or LLM embedding. Access control relies primarily on Kubernetes NetworkPolicies and cluster-level security controls since the protocol transport lacks internal authentication layers.

Categories