Is Docker required for this skill?

Docker Desktop is required if you are using 'Mode 1: Local PoC Setup' to host the local Grafana and Prometheus stack. It is not required for 'Enterprise Setup' if you are connecting to existing external infrastructure.

Do I need to restart Claude Code after setup?

Yes, any changes to the telemetry environment variables in settings.json require a full restart of Claude Code to take effect.

What specific metrics does this skill track?

It tracks session counts, active time, token usage (input, output, and cached), API costs in USD, lines of code modified, and Git actions like commits and PRs.

How do I access the monitoring dashboard?

In local mode, once the stack is started, you can access the pre-configured Grafana dashboard at http://localhost:3000 using the default credentials (admin/admin).

Can I connect this to my company's existing Datadog or Honeycomb?

Yes, by using the 'Enterprise Mode,' you can configure the OTEL exporter to point to any OTLP-compatible endpoint provided by your organization's observability platform.

Claude Code Monitoring & Metrics

Name: Claude Code Monitoring & Metrics
Author: cskiro

bycskiro

0•

分析と監視

Sets up comprehensive OpenTelemetry monitoring for tracking Claude Code usage, token costs, and developer productivity.

This skill automates the configuration of observability stacks for Claude Code, offering both a local Proof of Concept (PoC) mode via Docker and an Enterprise mode for centralized infrastructure. It enables teams and individual developers to visualize token consumption, calculate API costs by model, and track productivity metrics like lines of code modified and session activity through pre-configured Grafana dashboards. By integrating seamlessly with Prometheus and Grafana, it provides immediate visibility into the ROI and operational health of AI-assisted development workflows.

主な機能

01Automated local PoC stack deployment with Docker, Prometheus, and Grafana

02Productivity analytics including session time, LOC modified, and Git activity

030 GitHub stars

04Real-time tracking of token usage and associated USD costs by model

05Enterprise-ready configuration for centralized OpenTelemetry (OTLP) endpoints

06Built-in management scripts for starting, stopping, and cleaning up telemetry data

ユースケース

01Engineering managers monitoring the rollout and ROI of Claude Code across a team

02Individual developers wanting to track their personal AI spending and tool usage

03Enterprises requiring centralized audit logs and usage telemetry for compliance

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cskiro/claudex otel-monitoring-setup

For use in Claude.ai and ChatGPT

Download Skill