OpenShift OVN-Kubernetes Benchmark

Provides automated benchmarking, monitoring, and AI-powered analysis for OpenShift clusters using OVN-Kubernetes via an MCP server.

python

GitHub Stars

python

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "liqcui-ovnk-benchmark-mcp": {
      "command": "./ovnk_benchmark_mcp_command.sh",
      "args": [
        "-k",
        "~/.kube/config",
        "server"
      ],
      "env": {
        "KUBECONFIG": "~/.kube/config",
        "DATABASE_PATH": "storage/ovnk_benchmark.db",
        "MCP_SERVER_URL": "http://localhost:8000",
        "OPENAI_API_KEY": "YOUR_OPENAI_API_KEY",
        "REPORT_OUTPUT_DIR": "exports",
        "REPORT_PERIOD_DAYS": "7",
        "COLLECTION_DURATION": "5m"
      }
    }
  }
}

You run an MCP server that automates benchmarking, monitoring, and AI-assisted analysis for OpenShift clusters using OVN-Kubernetes networking. It collects metrics from multiple sources, analyzes trends, and generates actionable reports to help you optimize cluster performance.

How to use

You connect to your cluster through the MCP server client to start, collect data, and generate reports. Use the server to automatically authenticate with your OpenShift/Kubernetes cluster, gather metrics from Prometheus and the cluster API, and run AI-powered analyses that produce practical recommendations. Start by configuring access, then initiate data collection, review AI insights, and export executive summaries or full reports for stakeholders.

How to install

Prerequisites you need before installing and running the MCP server are clearly defined so you can prepare your environment.

# Prerequisites
python3.9+
OpenShift/Kubernetes cluster access
KUBECONFIG=~/.kube/config
OPENAI_API_KEY=YOUR_OPENAI_API_KEY

# Install and setup
git clone <repository>
cd ocp-benchmark-mcp
chmod +x ovnk_benchmark_mcp_command.sh
./ovnk_benchmark_mcp_command.sh setup

How to run and use the MCP server

Start the MCP server locally using the provided startup script, connecting to your cluster with your kubeconfig. This runs on port 8000 and will listen for MCP requests from your client.

# Start server (runs on port 8000)
./ovnk_benchmark_mcp_command.sh -k ~/.kube/config server

Collect data and generate reports

Collect performance data for a defined window, then generate a comprehensive report. You can also perform a full workflow to collect data and create a report in one step.

# Collect data for 10 minutes
./ovnk_benchmark_mcp_command.sh -k ~/.kube/config -d 10m collect

# Generate a report for the last 7 days (requires AI key for insights)
./ovnk_benchmark_mcp_command.sh -o YOUR_OPENAI_API_KEY -p 7 report

# Full workflow: collect data and generate a report
./ovnk_benchmark_mcp_command.sh -k ~/.kube/config -o YOUR_OPENAI_API_KEY full

Configuration and environment

Configure the MCP server with the recommended environment variables to enable data collection, AI analysis, and report output.

# Example environment setup
export KUBECONFIG=~/.kube/config
export OPENAI_API_KEY=YOUR_OPENAI_API_KEY
export MCP_SERVER_URL=http://localhost:8000
export COLLECTION_DURATION=5m
export REPORT_PERIOD_DAYS=7
export DATABASE_PATH=storage/ovnk_benchmark.db
export REPORT_OUTPUT_DIR=exports

# Optional: OpenShift Prometheus routing details if needed
export OVNK_PROMETHEUS_USE_ROUTE=true
export OVNK_PROMETHEUS_NAMESPACE=openshift-monitoring
export OVNK_PROMETHEUS_SA=prometheus-k8s

Notes and tips

The server uses a DuckDB store for historical performance data and generates Excel and PDF reports with AI-generated insights. Ensure your OpenShift cluster is reachable with the proper kubeconfig and that you provide a valid OpenAI API key if you want AI-driven analyses and recommendations.

Available tools

get_openshift_general_info

Fetch general cluster information including counts of NetworkPolicies, AdminNetworkPolicies, EgressFirewalls, namespaces, and nodes.

query_kube_api_metrics

Query Kubernetes API server performance metrics such as latency, duration, and error rates.

query_multus_metrics

Query Multus CNI performance metrics and resource usage.

query_ovnk_pods_metrics

Query OVN-Kubernetes pod performance metrics.

query_ovnk_containers_metrics

Query OVN-Kubernetes container metrics including database size and sync performance.

query_ovnk_sync_metrics

Query OVN-Kubernetes synchronization metrics.

store_performance_data

Store collected performance data into DuckDB storage.

get_performance_history

Retrieve historical performance data from storage for trend analysis.

test_connection

Test connectivity to the OpenShift cluster and MCP server health

Built by

VeilStrat

AI signals for GTM teams