Skip to content

usage.md

Latest commit

 

History

History
323 lines (247 loc) · 7 KB

usage.md

File metadata and controls

323 lines (247 loc) · 7 KB

Usage Information and Monitoring

Overview

Understanding your GitHub Copilot CLI usage is crucial for managing costs, optimizing performance, and staying within limits.

Basic Usage Commands

Check Current Usage

# Overall usage summary
gh copilot usage

# Detailed usage breakdown
gh copilot usage --detailed

# Usage for specific time periods
gh copilot usage --today
gh copilot usage --week
gh copilot usage --month

Context Window Information

# Current model's context window
gh copilot info context

# Context usage in current conversation
gh copilot info context --current

# Token count for specific text
gh copilot count-tokens "Your text here"
gh copilot count-tokens --file ./large-file.py

Understanding Usage Metrics

Token Usage

# Tokens used today
gh copilot usage tokens --today

# Token breakdown by model
gh copilot usage tokens --by-model

# Input vs output tokens
gh copilot usage tokens --split

Example Output:

Today's Usage:
┌─────────────┬─────────┬────────────┬─────────────┐
│ Model       │ Queries │ Input      │ Output      │
├─────────────┼─────────┼────────────┼─────────────┤
│ GPT-4       │ 15      │ 12,450     │ 3,200       │
│ GPT-3.5     │ 8       │ 2,100      │ 800         │
└─────────────┴─────────┴────────────┴─────────────┘

Request Limits

# Current rate limits
gh copilot limits

# Requests remaining
gh copilot limits --remaining

# Reset time for limits
gh copilot limits --reset-time

Cost Estimation

# Estimated costs for usage
gh copilot cost --today
gh copilot cost --week --detailed

# Cost by model
gh copilot cost --by-model

Context Window Management

Monitoring Context Usage

# Check current conversation token count
gh copilot context size

# Warning when approaching limit
gh copilot context check --warn-at 80%

# Detailed context breakdown
gh copilot context analyze

Context Window Sizes by Model

Model Context Window Recommended Usage
GPT-4 8,192 tokens ~6,000 tokens max
GPT-4-32K 32,768 tokens ~25,000 tokens max
GPT-3.5-Turbo 4,096 tokens ~3,000 tokens max
GPT-3.5-Turbo-16K 16,384 tokens ~12,000 tokens max

Managing Context Efficiently

# Clear context when full
gh copilot clear

# Start fresh conversation
gh copilot chat --new

# Summarize and continue
gh copilot summarize --continue

Advanced Usage Monitoring

Historical Usage

# Usage trends over time
gh copilot usage --history --last-30-days

# Export usage data
gh copilot usage --export csv --output usage-report.csv
gh copilot usage --export json --output usage-data.json

Team/Organization Usage

# Organization-wide usage (if admin)
gh copilot usage --org your-org

# Team usage summary
gh copilot usage --team your-team

# User breakdown
gh copilot usage --by-user

Performance Metrics

# Response time statistics
gh copilot stats response-time --last-week

# Model performance comparison
gh copilot stats models --compare

# Tool usage statistics
gh copilot stats tools --detailed

Setting Usage Alerts

Configure Warnings

# Set daily token limit warning
gh copilot config set alerts.daily_tokens 10000

# Set cost warning threshold
gh copilot config set alerts.daily_cost 5.00

# Context window warning
gh copilot config set alerts.context_usage 0.8

Alert Configuration

Edit ~/.config/gh/copilot/config.yml:

alerts:
  daily_tokens: 10000
  daily_cost: 5.00
  context_usage: 0.8
  rate_limit_threshold: 0.9
  notify:
    email: false
    desktop: true
    terminal: true

Usage Notifications

# Enable desktop notifications
gh copilot config set notifications.desktop true

# Email alerts (requires setup)
gh copilot config set notifications.email your-email@domain.com

# Slack integration (if configured)
gh copilot config set notifications.slack "#copilot-alerts"

Optimization Strategies

Token Efficiency

# Analyze token usage patterns
gh copilot analyze usage --patterns

# Suggestions for optimization
gh copilot suggest optimization

Best Practices:

  1. Use precise questions instead of broad ones
  2. Include only relevant context files
  3. Clear conversation when changing topics
  4. Use appropriate models for task complexity
  5. Batch related questions in one conversation

Cost Management

# Set spending limits
gh copilot config set limits.daily_cost 10.00
gh copilot config set limits.monthly_cost 100.00

# Auto-switch to cheaper models when near limit
gh copilot config set auto_optimize.cost true

# Pause usage when limit reached
gh copilot config set limits.auto_pause true

Model Optimization

# Automatic model selection based on complexity
gh copilot config set auto_optimize.model true

# Model switching based on context size
gh copilot config set auto_optimize.context true

# Performance-based model selection
gh copilot config set auto_optimize.performance true

Usage Reports

Daily Reports

# Generate daily summary
gh copilot report daily --date 2024-01-15

# Email daily report
gh copilot report daily --email

Weekly/Monthly Reports

# Weekly usage summary
gh copilot report weekly

# Monthly detailed report
gh copilot report monthly --detailed --export pdf

Custom Reports

# Custom date range
gh copilot report --from 2024-01-01 --to 2024-01-31

# Specific metrics
gh copilot report --metrics tokens,cost,response_time

# Compare periods
gh copilot report --compare --last-month --this-month

API and Integration

Usage API

# Get usage data via API
gh copilot api usage --json

# Historical data
gh copilot api usage --from "2024-01-01" --to "2024-01-31"

# Real-time monitoring
gh copilot api usage --stream

Webhook Integration

# Setup usage webhooks
gh copilot webhook create --url "https://your-domain.com/webhook" --events usage

# Configure webhook alerts
gh copilot webhook config --threshold-exceeded --daily-limit

Troubleshooting Usage Issues

Common Problems

Unexpected High Usage:

gh copilot usage --detailed --break-by conversation
gh copilot audit --large-requests --last-week

Context Window Errors:

gh copilot context diagnose
gh copilot context optimize --auto

Rate Limit Issues:

gh copilot limits check
gh copilot queue status

Recovery Commands

# Reset usage counters (if allowed)
gh copilot usage reset --confirm

# Clear all conversations to free up quota
gh copilot clear --all --confirm

# Switch to free tier model temporarily
gh copilot config set model gpt-3.5-turbo --temporary