talkopsai/opentelemetry-mcp-server repository overview

⁠OpenTelemetry MCP Server by TalkOps.ai⁠

An MCP server that gives AI assistants the power to discover, provision, instrument, validate, and govern OpenTelemetry pipelines on Kubernetes — from intent-driven collector provisioning to cardinality control, sampling optimization, and security auditing.

⁠✨ Features

• Intent-Driven Collector Provisioning: Express what you need (signals, namespace) — the tool auto-discovers backends (Jaeger, Tempo, Prometheus, Loki, and 6 more), generates best-practice configs with correct processor ordering, selects deployment mode, sizes resources from cluster scale, and creates RBAC automatically.
• Zero-to-Instrumented Onboarding: Look up language support, create Instrumentation CRDs, annotate Deployments for auto-instrumentation injection, and verify the rollout — all with dry-run-first safety.
• Pipeline Investigation & Validation: Deep-inspect any OTel Collector's config, validate processor ordering against best practices, audit filelog receiver safety, and check k8sattributes enrichment profiles.
• Metric Cardinality Governance: Detect high-cardinality dimensions from SpanMetrics connectors, generate transform processor YAML to drop attributes, and estimate series counts before they explode.
• Sampling Strategy Optimization: Cross-reference head sampling (Instrumentation CRDs) with tail sampling (collector config), detect conflicts, and generate config patches to switch strategies.
• Security Posture Auditing: Scan eBPF instrumentation pods for privileged mode, SYS_ADMIN capabilities, and hostPID access. Risk-assess the entire observability footprint.

⁠🚀 How to Use This Image

This Docker image runs an MCP server that connects to your Kubernetes cluster to manage OpenTelemetry resources. AI assistants (Claude, Cline, Cursor, or any MCP client) connect to the server over HTTP, SSE, or stdio.

⁠Running Standalone (HTTP Transport)

docker run --rm -it \
  -p 8771:8771 \
  -v ~/.kube:/app/.kube:ro \
  -e K8S_KUBECONFIG=/app/.kube/config \
  -e MCP_TRANSPORT=http \
  talkopsai/opentelemetry-mcp-server:latest

The server is now listening on http://localhost:8771/mcp.

Then configure your MCP client to connect:

{
  "mcpServers": {
    "opentelemetry": {
      "url": "http://localhost:8771/mcp",
      "description": "MCP Server for OpenTelemetry Kubernetes observability"
    }
  }
}

⁠In-Cluster Deployment

When running inside a Kubernetes pod, use in-cluster config:

docker run --rm -it \
  -p 8771:8771 \
  -e MCP_TRANSPORT=http \
  -e K8S_IN_CLUSTER=true \
  talkopsai/opentelemetry-mcp-server:latest

⁠Important: RBAC Requirements

The OpenTelemetry Operator must be installed in the cluster. The server's service account needs:

• get, list, watch on OpenTelemetryCollector and Instrumentation CRDs
• create, patch on OpenTelemetryCollector and Instrumentation CRDs (for provisioning)
• create, patch, get on ClusterRole and ClusterRoleBinding (for k8sattributes RBAC)

⁠🛠️ Configuration (Claude Desktop — Stdio)

To use this image over stdio transport directly inside Claude Desktop:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "opentelemetry-mcp-server": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-v", "/Users/YOUR_USERNAME/.kube:/app/.kube:ro",
        "-e", "K8S_KUBECONFIG=/app/.kube/config",
        "-e", "MCP_TRANSPORT=stdio",
        "-e", "MCP_LOG_LEVEL=INFO",
        "talkopsai/opentelemetry-mcp-server:latest"
      ]
    }
  }
}

Replace /Users/YOUR_USERNAME/.kube with your actual kubeconfig path (e.g. C:\Users\YourUser\.kube on Windows).

⁠⚙️ Environment Variables

⁠Server Configuration

⁠Kubernetes

⁠OTel Operator CRD

⁠Prometheus Integration

⁠🔧 Available Tools (19 Total)

⁠Discovery

⁠Collector Management

⁠Instrumentation

⁠Validation

⁠Governance

⁠Sampling

⁠SpanMetrics

⁠📖 Available Resources (9 URIs)

⁠🔐 Security Best Practices

1. Never expose the MCP server to the public internet without proper authentication.
2. Read-Only Mount: Always mount kubeconfig with :ro so the container cannot modify your credentials.
3. RBAC Least Privilege: Ensure the kubeconfig identity has only the permissions needed for OTel CRD management.
4. Dry-Run First: All mutating tools (otel_provision_collector, otel_patch_collector, otel_patch_instrumentation, otel_annotate_deployment) default to dry_run=True. Review output before applying.
5. Review RBAC Resources: otel_provision_collector creates ClusterRole/ClusterRoleBinding for k8sattributes — review the spec before applying with dry_run=False.

⁠🐳 Docker Compose Example

services:
  opentelemetry-mcp-server:
    image: talkopsai/opentelemetry-mcp-server:latest
    ports:
      - "8771:8771"
    environment:
      - MCP_TRANSPORT=http
      - K8S_IN_CLUSTER=true
      - MCP_LOG_LEVEL=INFO

⁠🔗 Links and Resources

• Source Code: GitHub Repository⁠
• Full Documentation: README & Guides⁠
• Model Context Protocol: MCP Official Site⁠
• OpenTelemetry: OpenTelemetry Documentation⁠
• OpenTelemetry Operator: GitHub Repository⁠
• Report Issues: GitHub Issues⁠
• Discord: Join the Community⁠
• License: Apache 2.0⁠

If you find this MCP server useful, consider leaving a ⭐ on the GitHub repository⁠!