Kentik MCP Server
A Model Context Protocol (MCP) server that provides LLM access to the Kentik network observability platform.
Disclaimer: This is an unofficial, community-maintained project and is not affiliated with, endorsed by, or supported by Kentik, Inc. "Kentik" is a trademark of Kentik, Inc. This project uses the name solely to describe its compatibility with Kentik's publicly documented APIs.
Features
Prerequisites
- Go 1.21+
- A Kentik account with API access
- Your Kentik email and API token (found in User Profile → Authentication)
Installation
go install github.com/awlx/kentik-mcp@latest
Or build from source:
git clone https://github.com/awlx/kentik-mcp.git
cd kentik-mcp
go build -o kentik-mcp .
Configuration
Set the following environment variables:
export [email protected]
export KENTIK_API_TOKEN=your_api_token_here
export KENTIK_REGION=US
Usage
With VS Code / GitHub Copilot
Add to your .vscode/mcp.json:
{
"servers": {
"kentik": {
"command": "kentik-mcp",
"env": {
"KENTIK_EMAIL": "[email protected]",
"KENTIK_API_TOKEN": "your_api_token_here",
"KENTIK_REGION": "US"
}
}
}
}
With Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"kentik": {
"command": "kentik-mcp",
"env": {
"KENTIK_EMAIL": "[email protected]",
"KENTIK_API_TOKEN": "your_api_token_here",
"KENTIK_REGION": "US"
}
}
}
}
Standalone (stdio)
./kentik-mcp
The server communicates over stdio using the MCP protocol.
Example Queries
Once connected, you can ask your LLM things like:
- "Search for border routers in the NYC site"
- "Show me the top 10 destination ASNs by traffic on PNI links in the last 24 hours"
- "What's the traffic breakdown by port on external links?"
- "Query flows per second by destination port, filtered to transit links"
- "What synthetic tests are configured?"
- "Ask Kentik AI: How are my devices doing?"
- "Which sites have the most devices?"
API Coverage
This MCP server covers:
- V5 REST APIs: Devices, interfaces, users, sites, labels, tags, and flow data queries
- V6 gRPC-gateway APIs: Synthetic monitoring (tests, agents, results, traces) and AI Advisor
For full Kentik API documentation, see: https://kb.kentik.com/docs/apis-overview
Rate Limits
Kentik applies rate limits per customer. Key limits:
AI Advisor has additional limits: 4 requests/min for create/update, 60 requests/min for polling.
License
MIT — see LICENSE.
This project is not affiliated with Kentik, Inc.