gemini-cli-health-dashboard

🩺 Gemini CLI Health Dashboard

This repository contains the JSON configuration and setup guide for the Gemini CLI Health Dashboard.

This dashboard provides real-time observability into the Gemini CLI, allowing Platform Engineering and DevOps teams to monitor API health, success vs. error rates, and response code distributions (2xx, 4xx, 5xx).

🛠️ Prerequisites

Before importing the dashboard, you must configure the Gemini CLI to send telemetry to Google Cloud and create specific Log-Based Metrics. The dashboard relies on these metrics to populate the charts.

1. 📡 Enable Gemini CLI Telemetry

Update your Gemini CLI settings file (typically ~/.gemini/settings.json) to enable telemetry and target Google Cloud:

{
  "telemetry": {
    "enabled": true,
    "target": "gcp"
  }
}

2. 📊 Create Log-Based Metrics

You must create 3 user-defined metrics in Google Cloud Logging using either option A or B .

⚠️ Important: The metric names must match exactly as listed below, or the dashboard widgets will show "No Data." Also replace project_id variable with your respective google cloud project ID.

[A] Setup metrics from Google Cloud Console

•       🟢 Metric A: Success Rate

  • Metric Type: Counter
  • Log Metric Name: gemini_cli_success
  • Labels:

    Name : success_status
    Type : STRING
    Field Name : jsonPayload."event.name"
    

  • Filter:

    logName="projects/<project_id>/logs/gemini_cli"
    jsonPayload.status_code:*
    jsonPayload."event.name"="gemini_cli.api_response"
    

•       🔴 Metric B: Error Rate

  • Metric Type: Counter
  • Log Metric Name: gemini_cli_error
  • Labels:

    Name : error_status
    Type : STRING
    Field Name : jsonPayload."event.name"
    

  • Filter:

    logName="projects/<project_id>/logs/gemini_cli"
    jsonPayload.status_code:*
    jsonPayload."event.name"="gemini_cli.api_error"
    

•       📈 Metric C: Response Distribution

  • Metric Type: Counter
  • Log Metric Name: gemini_cli_response
  • Labels: Create 2 labels

    Name : status
    Type : STRING
    Field Name : jsonPayload.status_code
    
    Name : event_name
    Type : STRING
    Field Name : jsonPayload."event.name"
    

  • Filter:

    logName="projects/<project_id>/logs/gemini_cli"
    jsonPayload.status_code:*
    

[B] Setup metrics using python script (Automated)

Instead of manually creating metrics, you can also run the provided script:

1. Open the terminal.
2. Authenticate using Google Cloud Credentials
3. Setup Google Cloud Project if not already set.
4. Clone this git repository.
5. Navigate path gemini-cli-health-dashboard/scripts/
6. Run: python3 create_log_based_metrics.py

📥 Installation

1. Download the gemini-cli-dashboard.json file from this repository.
2. Go to the Google Cloud Console > Monitoring > Dashboards.
3. Click Create Dashboard.
4. Switch to the JSON Editor tab.
5. Copy and paste the contents of gemini-cli-dashboard.json.
6. Click Apply.

🖼️ Dashboard Panels

• [A] Success vs Error Rate: 📊 A stacked bar chart showing the ratio of successful requests to failed ones over time.
• [B] Response Code Distribution: 📉 A line chart visualizing the volume of traffic segmented by HTTP response codes (200s, 400s, 500s).

🤝 Contributing

Feel free to submit Pull Requests to improve the dashboard layout or add new metrics!