PTP MCP Server
OfficialClick on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@PTP MCP Servercheck PTP health status"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
PTP MCP Server
A Model Context Protocol (MCP) server for monitoring and analyzing Precision Time Protocol (PTP) systems in OpenShift clusters.
๐ Features
PTP Configuration Analysis: Parse and validate PTP configurations from OpenShift
Real-time Log Monitoring: Access linuxptp daemon logs with intelligent parsing
Natural Language Queries: Ask questions about PTP status in plain English
Health Monitoring: Comprehensive PTP system health checks
Synchronization Analysis: Monitor sync status, offsets, and BMCA state
Clock Hierarchy: Track grandmaster and clock hierarchy information
ITU-T Compliance: Validate configurations against ITU-T G.8275.1 standards
๐ Prerequisites
Python 3.8 or higher
OpenShift CLI (
oc) installed and configuredAccess to OpenShift cluster with PTP operator installed
PTP namespace (
openshift-ptp) exists
๐ ๏ธ Source Installation
Clone the repository:
git clone https://github.com/redhat-cne/ptp-mcp-server.git cd ptp-mcp-server
To Deploy For Quick Tests
Install dependencies:
pip install -r requirements.txtVerify OpenShift access:
export KUBECONFIG=/path/to/kubeconfig oc whoami oc get namespace openshift-ptpQuick Testing
Run the comprehensive test suite:
export KUBECONFIG=/path/to/kubeconfig
python quick_test.pyExpected output:
๐ PTP MCP Server API Quick Test
==================================================
Tests Passed: 8/8
Success Rate: 100.0%
๐ ALL TESTS PASSED! Your API is ready for agent integration.๐ง MCP Server
The MCP server supports two transport modes:
stdio: For local MCP clients (Claude Code, Claude Desktop)
HTTP/SSE: For OpenShift Lightspeed integration
Local Usage (stdio mode)
python ptp_mcp_server.pyRemote Usage (http mode)
# Default port 8080
python ptp_mcp_server.py --http
# Custom port
python ptp_mcp_server.py --http --port 9000
# Or use environment variable
PTP_MCP_PORT=9000 python ptp_mcp_server.py --httpDeploy to OpenShift
podman build -t quay.io/$USER/ptp-mcp-server:latest .
podman push quay.io/$USER/ptp-mcp-server:latest
cd k8s && kustomize edit set image quay.io/redhat-cne/ptp-mcp-server=quay.io/$USER/ptp-mcp-server:latest && cd ..
oc apply -k k8s/Configure in OpenShift Lightspeed
Add the MCP server to your OLSConfig:
apiVersion: ols.openshift.io/v1alpha1
kind: OLSConfig
metadata:
name: cluster
spec:
featureGates:
- MCPServer
mcpServers:
- name: ptp-monitoring
url: 'http://ptp-mcp-server.openshift-ptp.svc.cluster.local:8080/mcp'
timeout: 30Usage in OpenShift Lightspeed
By default, the MCP server assumes that it is running on the cluster that is to be monitored/queried. Any tool use will target the local cluster. However, if the MCP server is running on a hub cluster and the user intends for its prompt to target spoke cluster then OLS context should include a base64 copy of the kubeconfig for the spoke cluster.
note: It is important to use a minimal kubeconfig that is token based rather than a client certificate based kubeconfig as the OLS context size will not allow for a larger kubeconfig value.
The following example generates such a kubeconfig for lab testing purposes. In a production environment care should be taken to limit the service account permissions to only the strict minimum RBAC policies required.
oc create sa ols-ptp-user -n default 2>/dev/null
oc adm policy add-cluster-role-to-user cluster-admin -z ols-ptp-user -n default
TOKEN=$(oc create token ols-ptp-user -n default --duration=24h)
API_SERVER=$(oc whoami --show-server)
cat << EOF > minimal-kubeconfig.yaml
apiVersion: v1
kind: Config
clusters:
- cluster:
server: ${API_SERVER}
insecure-skip-tls-verify: true
name: cluster
contexts:
- context:
cluster: cluster
user: user
name: ctx
current-context: ctx
users:
- name: user
user:
token: ${TOKEN}
EOF๐ API Endpoints
1. Configuration API
from ptp_tools import PTPTools
tools = PTPTools()
result = await tools.get_ptp_config({"namespace": "openshift-ptp"})2. Logs API
result = await tools.get_ptp_logs({"lines": 1000})3. Search API
result = await tools.search_logs({"query": "dpll", "time_range": "last_hour"})4. Health API
result = await tools.check_ptp_health({"check_config": True, "check_sync": True})5. Natural Language API
result = await tools.query_ptp({"question": "What is the current grandmaster?"})6. Grandmaster Status API
result = await tools.get_grandmaster_status({"detailed": True})7. Sync Status API
result = await tools.analyze_sync_status({"include_offsets": True})8. Clock Hierarchy API
result = await tools.get_clock_hierarchy({"include_ports": True})๐ Usage Examples
Basic Health Check
import asyncio
from ptp_tools import PTPTools
async def check_health():
tools = PTPTools()
health = await tools.check_ptp_health({})
if health["success"]:
print(f"Status: {health['overall_status']}")
for check_name, result in health["checks"].items():
print(f"{check_name}: {result}")
else:
print(f"Error: {health.get('error')}")
asyncio.run(check_health())Natural Language Query
async def ask_question():
tools = PTPTools()
response = await tools.query_ptp({
"question": "What is the current grandmaster?"
})
if response["success"]:
print(f"Answer: {response['response']}")
else:
print(f"Error: {response.get('error')}")
asyncio.run(ask_question())Log Analysis
async def analyze_logs():
tools = PTPTools()
# Get recent logs
logs = await tools.get_ptp_logs({"lines": 500})
# Search for specific events
sync_loss = await tools.search_logs({"query": "sync loss"})
clock_changes = await tools.search_logs({"query": "clockClass change"})
print(f"Total logs: {logs['logs_count']}")
print(f"Sync loss events: {sync_loss['matching_logs']}")
print(f"Clock changes: {clock_changes['matching_logs']}")
asyncio.run(analyze_logs())MCP Tools Available
Tool | Description |
| Get PTP configuration |
| Get linuxptp daemon logs |
| Search logs for patterns |
| Get grandmaster info |
| Analyze sync status |
| Get clock hierarchy |
| Comprehensive health check |
| Natural language interface |
| Execute PMC commands |
Deployment Files
File | Purpose |
| Container image with Python + oc CLI |
| ServiceAccount, ClusterRole, ClusterRoleBinding |
| Deployment with health checks |
| ClusterIP Service |
| Example OLS configuration |
| Deploy all with |
RBAC Permissions
The ServiceAccount is granted these permissions:
Resource | Verbs | Purpose |
| get, list, watch | Read PTP configurations |
| get, list, watch | Find linuxptp-daemon pods |
| get, list | Read daemon logs |
| create | Execute PMC queries |
| get, list | Namespace access |
| get, list | Node topology (optional) |
๐ Performance
Average Response Time: 0.78s
Fastest API: Configuration API (0.22s)
Concurrent Operations: 4/4 successful in 2.45s
Success Rate: 100% (8/8 endpoints)
๐๏ธ Architecture
ptp-mcp-server/
โโโ ptp_mcp_server.py # Main MCP server (stdio + HTTP modes)
โโโ ptp_config_parser.py # PTP configuration parser
โโโ ptp_log_parser.py # Linuxptp log parser
โโโ ptp_model.py # PTP data models
โโโ ptp_query_engine.py # Natural language query engine
โโโ ptp_tools.py # API endpoint implementations
โโโ quick_test.py # Quick test suite
โโโ performance_test.py # Performance benchmarking
โโโ requirements.txt # Python dependencies
โโโ Dockerfile # Container image definition
โโโ k8s/ # Kubernetes/OpenShift manifests
โโโ kustomization.yaml # Kustomize configuration
โโโ rbac.yaml # ServiceAccount & RBAC
โโโ deployment.yaml # Deployment specification
โโโ service.yaml # Service definition
โโโ olsconfig-example.yaml # OLS integration example๐ PTP Concepts Supported
BMCA (Best Master Clock Algorithm): Clock selection and hierarchy
Clock Types: OC (Ordinary Clock), BC (Boundary Clock), TC (Transparent Clock)
ITU-T G.8275.1: Profile compliance and validation
Synchronization: Offset tracking, frequency adjustment, sync status
Grandmaster: Primary time source identification and status
Clock Class: Quality and traceability indicators
Domain Numbers: PTP domain configuration (24-43 for ITU-T)
๐งช Testing
Run All Tests
python quick_test.pyPerformance Testing
python performance_test.pyIndividual Component Testing
# Test configuration parser
python -c "from ptp_config_parser import PTPConfigParser; import asyncio; asyncio.run(PTPConfigParser().get_ptp_configs())"
# Test log parser
python -c "from ptp_log_parser import PTPLogParser; import asyncio; asyncio.run(PTPLogParser().get_ptp_logs())"๐ Documentation
Testing Guide - Comprehensive testing instructions
Agent Integration Guide - Integration examples for agents
Testing Steps - Step-by-step testing process
Testing Results - Complete test results
๐ค Contributing
Fork the repository
Create a feature branch (
git checkout -b feature/amazing-feature)Commit your changes (
git commit -m 'Add amazing feature')Push to the branch (
git push origin feature/amazing-feature)Open a Pull Request
๐ License
This project is licensed under the MIT License - see the LICENSE file for details.
๐ Acknowledgments
OpenShift PTP Operator team
Linuxptp project
Model Context Protocol (MCP) community
๐ Support
For issues and questions:
Create an issue on GitHub
Check the testing documentation
Review the agent integration guide
Status: โ
Production Ready
Last Updated: January 2025
Version: 1.0.0
This server cannot be installed
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/redhat-cne/ptp-mcp-server'
If you have feedback or need assistance with the MCP directory API, please join our Discord server