Can I track my actual Claude API spending with this skill?

Yes, the skill includes a pre-built Cost Analysis dashboard that tracks token usage and provides cost projections based on your session activity.

Does this require a Grafana Cloud account?

No, this skill deploys a fully self-hosted stack locally using Docker Compose, ensuring your data remains on your machine.

Is my prompt data kept private during monitoring?

This setup logs full prompt content and tool details to your local Loki instance. Because the stack runs on your local infrastructure, you maintain total control over your data privacy.

What services are included in this observability stack?

The stack includes Grafana for visualization, Loki for logs, Tempo for distributed tracing, Prometheus for metrics, and Grafana Alloy as the telemetry receiver.

How long does the setup process take?

The automated deployment typically takes between 5 to 10 minutes, including the time to download Docker images and generate configurations.

Claude Observability Stack Setup

Name: Claude Observability Stack Setup
Author: adaptationio

byadaptationio

0•

分析与监控

Deploys a comprehensive LGTM and Grafana Alloy observability stack via Docker Compose to monitor Claude Code performance, costs, and logs.

This skill automates the deployment of a professional-grade observability infrastructure specifically tailored for monitoring Claude Code operations. It installs and configures the 'LGTM' stack—consisting of Loki for log aggregation, Grafana for visualization, Tempo for distributed tracing, and Prometheus for metrics—alongside Grafana Alloy as the OTLP collector. By providing pre-configured dashboards for token usage, tool performance, and error tracking, it allows developers to analyze session costs, debug agentic tool calls, and optimize their local AI development workflow with deep telemetry insights.

主要功能

01Pre-configured Grafana dashboards for Claude Code cost and performance analysis

02Automated service health checks and infrastructure verification

03Integrated Grafana Alloy OTLP collector for seamless telemetry ingestion

040 GitHub stars

05One-command deployment of the full LGTM observability stack

06Persistent local storage for long-term historical data analysis

使用场景

01Creating local development environments with enterprise-standard monitoring

02Tracking Claude Code token consumption and projecting monthly API costs

03Debugging complex tool call failures and latency issues in real-time

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add adaptationio/skrillz observability-stack-setup

For use in Claude.ai and ChatGPT

Download Skill