
OpenShift OVN-Kubernetes Benchmark MCP Server
Enables comprehensive benchmarking and performance monitoring of OpenShift clusters using OVN-Kubernetes networking through automated data collection, AI-powered analysis, and report generation. Provides intelligent insights into cluster performance, bottleneck detection, and optimization recommendations.
README
OpenShift OVN-Kubernetes Benchmark MCP Server
A comprehensive benchmarking and performance monitoring solution for OpenShift clusters using OVN-Kubernetes networking, built with FastMCP and AI-powered analysis.
Architecture Overview
Architecture Topology
┌─────────────────────────────────────────────────────────────────────────────────┐
│ OpenShift Cluster │
│ ┌──────────────────┐ ┌──────────────────┐ ┌──────────────────┐ │
│ │ Kubernetes │ │ Prometheus │ │ OVN-Kubernetes │ │
│ │ API │ │ Metrics │ │ Components │ │
│ └──────────────────┘ └──────────────────┘ └──────────────────┘ │
└─────────────────────────────────────────────────────────────────────────────────┘
│
│ KUBECONFIG + SA Token
│
┌─────────────────────────────────────────────────────────────────────────────────┐
│ MCP Server Application │
│ │
│ ┌──────────────────┐ ┌──────────────────┐ ┌──────────────────┐ │
│ │ Authentication │ │ Data Collection│ │ Performance │ │
│ │ Module │ │ Tools │ │ Analysis │ │
│ │ │ │ │ │ │ │
│ │ • Auth Manager │ │ • Cluster Info │ │ • Bottleneck │ │
│ │ • Token Mgmt │ │ • Prometheus │ │ Detection │ │
│ │ │ │ Queries │ │ • Trend Analysis│ │
│ └──────────────────┘ └──────────────────┘ └──────────────────┘ │
│ │ │
│ ┌──────────────────┐ ┌──────────────────┐ ┌──────────────────┐ │
│ │ Data Storage │ │ ETL Processing │ │ Report Gen. │ │
│ │ │ │ │ │ │ │
│ │ • DuckDB │ │ • JSON to Table │ │ • Excel Reports │ │
│ │ • Time Series │ │ • Data Transform│ │ • PDF Summary │ │
│ │ Storage │ │ • Aggregation │ │ • HTML Dashboard│ │
│ └──────────────────┘ └──────────────────┘ └──────────────────┘ │
│ │
│ ┌─────────────────────────────────────────────────────────────────────────────┐│
│ │ AI Agents (LangGraph) ││
│ │ ││
│ │ ┌──────────────────┐ ┌──────────────────┐ ││
│ │ │ Performance │ │ Report │ ││
│ │ │ Data Agent │ │ Generation │ ││
│ │ │ │ │ Agent │ ││
│ │ │ • Collect Metrics│ │ • Analyze Data │ ││
│ │ │ • Store in DB │ │ • Compare History│ ││
│ │ │ • Real-time Mon. │ │ • Generate Report│ ││
│ │ └──────────────────┘ └──────────────────┘ ││
│ └─────────────────────────────────────────────────────────────────────────────┘│
└─────────────────────────────────────────────────────────────────────────────────┘
│
│ MCP Protocol (StreamableHTTP)
│
┌─────────────────────────────────────────────────────────────────────────────────┐
│ MCP Client Chat/API │
│ (Claude/LLM Interface) │
└─────────────────────────────────────────────────────────────────────────────────┘
graph TB
subgraph "OpenShift Cluster"
OCP[OpenShift Cluster]
PROM[Prometheus]
KAPI[Kubernetes API]
OVNK[OVN-Kubernetes Pods]
MULTUS[Multus CNI]
end
subgraph "MCP Client Layer"
CLIENT_API[MCP Client API<br/>REST/WebSocket Interface]
CLIENT_CHAT[MCP Client Chat<br/>with LLM Integration]
end
subgraph "MCP Server Layer"
MCP[FastMCP Server<br/>Port 8000]
AUTH[Authentication Module]
TOOLS[MCP Tools]
end
subgraph "AI Agents"
AGENT1[Performance Data<br/>Collection Agent]
AGENT2[Report Generation<br/>Agent]
LLM[OpenAI GPT-4]
end
subgraph "Storage & Reports"
DUCK[DuckDB Storage]
EXCEL[Excel Reports]
PDF[PDF Reports]
end
%% OpenShift to Server connections
OCP --> AUTH
PROM --> TOOLS
KAPI --> TOOLS
%% Client to Server connections
CLIENT_API --> MCP
CLIENT_CHAT --> MCP
CLIENT_CHAT --> LLM
%% Server internal connections
AUTH --> MCP
TOOLS --> MCP
%% Agent connections
AGENT1 --> MCP
AGENT2 --> MCP
AGENT1 --> LLM
AGENT2 --> LLM
%% Storage connections
MCP --> DUCK
AGENT2 --> EXCEL
AGENT2 --> PDF
%% Styling
style CLIENT_API fill:#fff3e0
style CLIENT_CHAT fill:#fff3e0
style MCP fill:#e1f5fe
style AGENT1 fill:#f3e5f5
style AGENT2 fill:#f3e5f5
style DUCK fill:#e8f5e8
Features
🔧 Core Capabilities
- Automated Authentication: Discovers and authenticates with OpenShift/Kubernetes clusters
- Multi-Source Monitoring: Collects metrics from Prometheus, Kubernetes API, and cluster resources
- AI-Powered Analysis: Uses LangGraph and OpenAI for intelligent insights and recommendations
- Comprehensive Reporting: Generates Excel and PDF reports with visualizations
- Historical Tracking: Stores performance data in DuckDB for trend analysis
📊 Monitored Components
- Kubernetes API Server: Request latency, throughput, and error rates
- Multus CNI: Resource usage and pod networking performance
- OVN-Kubernetes Pods: Control plane and node performance
- OVN Containers: Database sizes, memory usage, and sync performance
- OVS Components: CPU and memory usage of OVS processes
- General Cluster Info: NetworkPolicies, AdminNetworkPolicies, EgressFirewalls
🤖 AI Features
- Automated performance trend analysis
- Intelligent alert correlation
- Proactive recommendations
- Risk assessment and health scoring
- Natural language insights
Quick Start
Prerequisites
- Python 3.9+
- OpenShift/Kubernetes cluster access
- KUBECONFIG file
- OpenAI API key (for AI features)
Installation
-
Clone and Setup
git clone <repository> cd ocp-benchmark-mcp chmod +x ovnk_benchmark_mcp_command.sh ./ovnk_benchmark_mcp_command.sh setup
-
Test Configuration
./ovnk_benchmark_mcp_command.sh -k ~/.kube/config test
Usage
Start MCP Server
# Start server (runs on port 8000)
./ovnk_benchmark_mcp_command.sh -k ~/.kube/config server
Collect Performance Data
# Collect data for 10 minutes
./ovnk_benchmark_mcp_command.sh -k ~/.kube/config -d 10m collect
Generate Reports
# Generate report for last 7 days
./ovnk_benchmark_mcp_command.sh -o sk-your-openai-key -p 7 report
Full Workflow
# Collect data and generate report
./ovnk_benchmark_mcp_command.sh -k ~/.kube/config -o sk-your-openai-key full
Project Structure
ocp-benchmark-mcp/
├── README.md # This file
├── pyproject.toml # Python project configuration
├── ovnk_benchmark_mcp_server.py # Main MCP server
├── ovnk_benchmark_mcp_agent_perfdata.py # Data collection agent
├── ovnk_benchmark_mcp_agent_report.py # Report generation agent
├── ovnk_benchmark_mcp_command.sh # Startup script
├── ocauth/
│ └── ovnk_benchmark_auth.py # OpenShift authentication
├── tools/
│ ├── ovnk_benchmark_openshift_generalinfo.py # Cluster general info
│ ├── ovnk_benchmark_prometheus_basequery.py # Base Prometheus queries
│ ├── ovnk_benchmark_prometheus_kubeapi.py # API server metrics
│ ├── ovnk_benchmark_prometheus_multus.py # Multus CNI metrics
│ ├── ovnk_benchmark_prometheus_ovnk_pods.py # OVN-K pod metrics
│ ├── ovnk_benchmark_prometheus_ovnk_containers.py # OVN container metrics
│ └── ovnk_benchmark_prometheus_ovnk_sync.py # OVN sync metrics
├── config/
│ ├── ovnk_benchmark_config.py # Configuration management
│ └── metrics.yml # Prometheus metrics definitions
├── analysis/
│ └── ovnk_benchmark_performance_analysis.py # Performance analysis
├── elt/
│ └── ovnk_benchmark_performance_elt.py # Data processing
├── storage/
│ └── ovnk_benchmark_prometheus_ovnk.py # DuckDB storage
└── exports/ # Generated reports
Configuration
Environment Variables
Variable | Description | Default |
---|---|---|
KUBECONFIG |
Path to kubeconfig file | ~/.kube/config |
OPENAI_API_KEY |
OpenAI API key for AI features | Required for reports |
MCP_SERVER_URL |
MCP server URL | http://localhost:8000 |
COLLECTION_DURATION |
Metrics collection duration | 5m |
REPORT_PERIOD_DAYS |
Report period in days | 7 |
DATABASE_PATH |
DuckDB database path | storage/ovnk_benchmark.db |
REPORT_OUTPUT_DIR |
Report output directory | exports |
Metrics Configuration
The config/metrics.yml
file defines all Prometheus queries organized by category:
- General Information: Pod and namespace status
- API Server: Request latency and error rates
- Multus: CNI resource usage
- OVN Control Plane/Node: CPU and memory metrics
- OVN Containers: Database and controller metrics
- OVS Containers: OVS daemon metrics
- OVN Sync: Synchronization duration metrics
API Reference
MCP Tools
The server exposes the following MCP tools:
get_openshift_general_info
Get general cluster information including NetworkPolicy, AdminNetworkPolicy, and EgressFirewall counts.
Parameters:
namespace
(optional): Specific namespace to query
Response:
{
"timestamp": "2024-01-01T00:00:00Z",
"summary": {
"total_networkpolicies": 10,
"total_adminnetworkpolicies": 2,
"total_egressfirewalls": 5,
"total_namespaces": 25,
"total_nodes": 6
}
}
query_kube_api_metrics
Query Kubernetes API server performance metrics.
Parameters:
duration
(optional): Query duration (default: "5m")start_time
(optional): Start time in ISO formatend_time
(optional): End time in ISO format
query_multus_metrics
Query Multus CNI performance metrics.
query_ovnk_pods_metrics
Query OVN-Kubernetes pod performance metrics.
query_ovnk_containers_metrics
Query OVN-Kubernetes container metrics.
query_ovnk_sync_metrics
Query OVN-Kubernetes synchronization metrics.
store_performance_data
Store performance data in DuckDB.
get_performance_history
Retrieve historical performance data.
AI Agents
Performance Data Collection Agent
Uses LangGraph to orchestrate data collection:
- Initialize: Setup collection parameters
- Collect General Info: Gather cluster information
- Collect Metrics: Query each component category
- Store Data: Save to DuckDB storage
- Finalize: Generate collection summary
Report Generation Agent
Uses LangGraph with AI analysis:
- Fetch Historical Data: Retrieve performance history
- Analyze Performance: Calculate trends and statistics
- Generate Insights: Use AI for recommendations
- Create Reports: Generate Excel and PDF reports
- Finalize: Output summary and files
Storage Schema
DuckDB Tables
metrics
: Individual metric data pointsmetric_summaries
: Category performance summariesperformance_snapshots
: Complete performance snapshotsbenchmark_runs
: Benchmark execution recordsalerts_history
: Historical alert data
Report Types
Excel Reports
- Executive Summary: Key performance indicators
- Historical Trends: Time-series performance data
- Category Analysis: Component-specific metrics
- Recommendations: AI-generated insights
- Raw Data: Complete dataset
PDF Reports
- Executive Summary: High-level performance overview
- Key Metrics: Performance indicator tables
- Category Analysis: Component performance breakdown
- Recommendations: Prioritized action items
Troubleshooting
Common Issues
Authentication Problems
# Test cluster connectivity
kubectl cluster-info
# Verify kubeconfig
export KUBECONFIG=/path/to/config
./ovnk_benchmark_mcp_command.sh test
Prometheus Discovery
# Check Prometheus pods
kubectl get pods -n openshift-monitoring | grep prometheus
# Verify service accounts
kubectl get sa -n openshift-monitoring
MCP Server Issues
# Check server logs
tail -f logs/mcp_server_*.log
# Test server connectivity
curl http://localhost:8000/health
Debug Mode
Enable debug logging:
export LOG_LEVEL=DEBUG
./ovnk_benchmark_mcp_command.sh server
Contributing
- Fork the repository
- Create a feature branch
- Add tests for new functionality
- Submit a pull request
Development Setup
# Install development dependencies
pip install -e .[dev]
# Run tests
pytest
# Format code
black .
# Type checking
mypy .
License
MIT License - see LICENSE file for details.
Support
For issues and questions:
- Check the troubleshooting section
- Review logs in the
logs/
directory - Open an issue with detailed logs and configuration
Roadmap
- [ ] Kubernetes native deployment (Helm charts)
- [ ] Grafana dashboard integration
- [ ] Custom alert rule definitions
- [ ] Multi-cluster support
- [ ] Real-time streaming metrics
- [ ] Advanced ML-based anomaly detection
- [ ] Integration with CI/CD pipelines
Note: This tool is designed for OpenShift clusters with OVN-Kubernetes networking. Some features may not be available on other Kubernetes distributions.
推荐服务器

Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。