What Is Claude Code Monitor?

Claude Code now exports rich telemetry data out of the box through OpenTelemetry — the official, vendor-neutral way to monitor everything your agentic coding sessions produce.

• Token usage (input, output, cache creation, cache read)
• Cost estimates (real-time and predictive)
• Tool activity (file edits, command execution, git operations, browser use)
• Background tasks & persistent sessions (/tasks, long-running agents)
• Productivity metrics (accepted code lines, suggestion acceptance rate)

Unlike earlier local JSONL-based community dashboards, official OTel support gives you enterprise-grade observability with zero custom code. Claude Code agents now run for hours or days autonomously — without monitoring, you risk surprise bills, lost progress, or rate-limit surprises.

// OpenTelemetry Collector data flow — the backbone of Claude Code's official monitoring architecture

// Real-time Claude Code session monitor (terminal + multi-session view)

Step-by-Step Setup Guide

Launch Claude Code with these environment variables. Add to your ~/.zshrc or ~/.bashrc for permanence:

export CLAUDE_CODE_ENABLE_TELEMETRY=1  export OTEL_METRICS_EXPORTER=otlp  export OTEL_LOGS_EXPORTER=otlp  export OTEL_TRACES_EXPORTER=otlp  export OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4317   # Point to your Collector  export OTEL_EXPORTER_OTLP_PROTOCOL=grpc  export OTEL_METRIC_EXPORT_INTERVAL=10000                    # 10 seconds for real-time feel

Restart your terminal and run claude — telemetry will now stream automatically.

Create a project folder and this docker-compose.yml:

version: '3.8'  services:    otel-collector:      image: otel/opentelemetry-collector-contrib:latest      command: ["--config", "/etc/otel/config.yaml"]      volumes:        - ./otel-config.yaml:/etc/otel/config.yaml      ports:        - "4317:4317"   # gRPC        - "4318:4318"   # HTTP      depends_on:        - prometheus      prometheus:      image: prom/prometheus:latest      ports:        - "9090:9090"      volumes:        - ./prometheus.yml:/etc/prometheus/prometheus.yml      grafana:      image: grafana/grafana:latest      ports:        - "3000:3000"      environment:        - GF_SECURITY_ADMIN_PASSWORD=admin      depends_on:        - prometheus

otel-config.yaml (minimal but powerful):

receivers:    otlp:      protocols:        grpc:        http:    processors:    batch:    exporters:    prometheus:      endpoint: "0.0.0.0:9464"    logging:    service:    pipelines:      metrics:        receivers: [otlp]        processors: [batch]        exporters: [prometheus]      logs:        receivers: [otlp]        processors: [batch]        exporters: [logging]

prometheus.yml:

scrape_configs:    - job_name: 'otel-collector'      static_configs:        - targets: ['otel-collector:9464']

docker compose up -d

1. Open http://localhost:3000 → login (admin/admin)
2. Add Prometheus data source (URL: http://prometheus:9090)
3. Import or create a dashboard with the key queries below

You'll instantly see beautiful graphs of token burn, cost curves, and agent health.

// Full observability pipeline: Claude Code → OTel Collector → Prometheus → Grafana

Key Metrics You Can Track Today

• Real-time token consumption with cache hit rate
• Cost forecasting — hours until you hit your limit
• Background Agent status (/tasks view + duration)
• Productivity ROI (accepted lines of code vs. time saved)
• Anomaly alerts (sudden token spikes = potential runaway agent)

Advanced: Cost Optimization + Multi-Model Strategy

Once you see exactly where your Claude Code spend is going, the next step is smart routing.

Pro tip: Use a unified OneAPI platform as a smart backend. Route complex reasoning → Claude Code (monitored via OTel). Route high-volume or repetitive tasks → cheaper models (Gemma 4, MAI, Hermes, etc.).

• Built-in usage logging & cost dashboards
• Automatic fallback & load balancing
• Zero-retention privacy options

Result: 30–60% lower costs while keeping full visibility.

Common Questions & Troubleshooting