Back to MCP Servers

Instana MCP Server

Official IBM Instana MCP server exposing unified tools that wrap Instana observability REST APIs. Supports Streamable HTTP and Stdio transports.

Observability by IBM Instana API Key active
Overview

The Instana MCP server is the official IBM-maintained bridge between AI agents and the Instana observability platform. It converts natural language requests into Instana REST API calls and returns structured responses, letting agents query application performance metrics, infrastructure topology, events, websites, SLOs, releases, dashboards, and automation actions in real time.

Rather than exposing a separate MCP tool for every Instana endpoint, the server consolidates the API surface into a small set of unified tools (such as manage_instana_resources, analyze_infrastructure, manage_events, and manage_slo) that internally route to the appropriate underlying endpoint. Infrastructure analysis is dynamically synchronized with the Instana API catalog so it can cover all entity types without a hard-coded tool list.

The server is written in Python, distributed via PyPI as mcp-instana, and licensed Apache-2.0. It supports two transports: Streamable HTTP (recommended, credentials passed per-request via headers) and Stdio (credentials passed via environment variables), with optional --tools filtering to enable only specific tool categories.

Tools

Tool Description
manage_instana_resources Application performance metrics, alert configurations, application settings, and catalog access.
analyze_infrastructure Two-pass infrastructure analysis with dynamic support for all entity types in the Instana API catalog.
manage_events Query Kubernetes events, agent monitoring events, and system event tracking.
manage_websites Website monitoring metrics, catalog browsing, configuration management, and analysis.
manage_automation Browse the automation action catalog and inspect action execution history.
manage_custom_dashboards CRUD operations on Instana custom dashboards.
manage_slo Manage SLO configurations, reports, alerts, and correction windows.
manage_releases Release tracking with pagination and filtering.
Setup Guide

Prerequisites

  • Python 3.x
  • An Instana tenant URL (for example https://your-instance.instana.io)
  • An Instana API token (or session tokens)
  • Node.js / npx if you use the mcp-remote proxy for Streamable HTTP mode

Install from PyPI

pip install mcp-instana

Run the server in Streamable HTTP mode (recommended):

mcp-instana --transport streamable-http

Or in Stdio mode:

mcp-instana --transport stdio

Claude Desktop config (Stdio)

{
  "mcpServers": {
    "Instana MCP Server": {
      "command": "mcp-instana",
      "args": ["--transport", "stdio"],
      "env": {
        "INSTANA_BASE_URL": "https://your-instance.instana.io",
        "INSTANA_API_TOKEN": "your_token"
      }
    }
  }
}

Claude Desktop config (Streamable HTTP via mcp-remote)

{
  "mcpServers": {
    "Instana MCP Server": {
      "command": "npx",
      "args": [
        "mcp-remote", "http://0.0.0.0:8080/mcp/",
        "--allow-http",
        "--header", "instana-base-url: https://your-instance.instana.io",
        "--header", "instana-api-token: your_token"
      ]
    }
  }
}

VS Code / GitHub Copilot (.vscode/mcp.json)

{
  "servers": {
    "Instana MCP Server": {
      "command": "mcp-instana",
      "args": ["--transport", "stdio"],
      "env": {
        "INSTANA_BASE_URL": "https://your-instance.instana.io",
        "INSTANA_API_TOKEN": "your_token"
      }
    }
  }
}

Optional tool filtering

Use --tools to enable only specific categories: router, infra, events, website, automation, dashboard, slo, releases.

Development install

pip install uv
uv sync
uv run src/core/server.py --transport streamable-http
Use Cases
  • Triage production incidents by asking an agent to pull recent Kubernetes events, agent health, and application alerts from Instana.
  • Analyze infrastructure across all monitored entity types without writing custom dashboards, using the dynamic catalog-aware analyze_infrastructure tool.
  • Track releases and correlate them with SLO breaches or error spikes for post-deploy validation.
  • Inspect and manage SLO configurations, reports, alerts, and correction windows directly from a chat interface.
  • Audit website performance by querying real user monitoring metrics and configuration for monitored sites.
Example Prompts
  • "Show me all Kubernetes events from the payments namespace in the last hour using Instana."
  • "Analyze the JVM heap usage across all hosts tagged production over the past 24 hours."
  • "List active SLO alerts and any open correction windows for the checkout service."
  • "Pull the latest 10 releases and flag any that coincided with error rate spikes in Instana."
  • "Create a custom Instana dashboard showing latency and error rate for the api-gateway application."
Pros
  • Officially maintained by IBM Instana under the instana GitHub org, Apache-2.0 licensed.
  • Wide coverage of the Instana platform: applications, infrastructure, events, websites, SLOs, releases, dashboards, and automation.
  • Supports both Stdio and Streamable HTTP transports, with per-request header credentials for multi-tenant or stateless deployments.
  • Dynamic entity-type support in analyze_infrastructure stays in sync with the Instana API catalog.
Limitations
  • Repository is still early-stage with no tagged releases and a small star count, so APIs and tool shapes may evolve.
  • Requires a valid Instana tenant and API token; there is no free public sandbox.
  • Streamable HTTP mode typically needs the mcp-remote proxy for clients like Claude Desktop, adding a moving part to the setup.
Alternatives