Usage Analytics provides real-time visibility into how your applications consume AI models across all providers, helping you optimize performance and identify trends.

​
Overview

Requesty’s Usage Analytics dashboard gives you complete visibility into your AI usage patterns across all models and providers. Track requests, tokens, latency, and user activity in real-time.

​
Why Usage Analytics Matters

Understanding your AI usage patterns is crucial for optimizing performance, controlling costs, and ensuring reliable service delivery.

​
Key Benefits

  • Identify Usage Trends: Spot patterns in request volumes to better plan capacity
  • Optimize Model Selection: See which models are used most and adjust your strategy
  • Track User Adoption: Monitor how users interact with AI features
  • Prevent Overages: Stay ahead of usage limits before they impact service
  • Improve Performance: Identify bottlenecks and optimization opportunities

​
Key Metrics

Request Volume

Monitor total requests across all models with hourly, daily, and monthly breakdowns

Token Usage

Track input, output, and cached tokens to understand consumption patterns

Response Times

Analyze P50, P90, and P99 latency metrics for performance optimization

User Activity

Monitor unique users and their usage patterns across different time periods

​
Dashboard Views

​
General Overview

The main dashboard provides at-a-glance metrics including:
  • Total Requests: Aggregate request count with trend indicators
  • Active Models: Number of unique models being used
  • Average Latency: Overall response time across all requests
  • Token Efficiency: Ratio of cached vs computed tokens

​
Time-Series Analysis

Visualize your usage patterns over time with customizable date ranges:
  • Hourly breakdown for the last 24 hours
  • Daily trends for the past 30 days
  • Monthly comparisons for year-over-year analysis
  • Custom date ranges for specific analysis needs

​
Model Breakdown

Detailed analytics per model including:
  • Request distribution across models
  • Token consumption by model type
  • Performance comparison between providers
  • Cost efficiency metrics per model

​
Filtering and Segmentation

​
Available Filters

  • Time Period: Last hour, day, week, month, or custom range
  • Model Provider: OpenAI, Anthropic, Google, etc.
  • Model Type: GPT-4, Claude, Gemini, etc.
  • API Keys: Filter by specific API keys
  • User Groups: Segment by user groups or teams
  • Request Status: Successful, failed, or cached requests

​
Group By Options

Organize your data by:
  • Model provider
  • Model name
  • API key
  • User or user group
  • Time bucket (hour, day, week)
  • Request status

​
Real-Time Monitoring

All analytics update in real-time with less than 1-second latency, ensuring you always have the most current data for decision-making.

​
Live Metrics

  • Active requests counter
  • Real-time token consumption rate
  • Current request queue size
  • Live error rate monitoring

​
Monitoring Best Practices

  • Check usage patterns daily during initial deployment
  • Review weekly trends to identify optimization opportunities
  • Monitor peak usage times to plan capacity
  • Track token efficiency to maximize cache utilization

​
Exporting and Reporting

​
Export Formats

Download your analytics data in multiple formats:
  • CSV: For spreadsheet analysis
  • JSON: For programmatic processing
  • PDF Reports: For stakeholder presentations

​
Scheduled Reports

Configure automated reports to be sent:
  • Daily summaries via email
  • Weekly performance reports
  • Monthly usage breakdowns
  • Custom reporting schedules

​
Data Export

​
Export Options

Download your usage data for further analysis:
  • CSV Format: Export to spreadsheets for custom analysis
  • JSON Format: Machine-readable format for processing
  • PDF Reports: Professional reports for stakeholders

​
Custom Reporting

Create custom views by:
  • Filtering by specific date ranges
  • Grouping by models, users, or API keys
  • Selecting specific metrics to export
  • Scheduling regular report generation

​
Best Practices

​
Integration with Other Features

Usage Analytics seamlessly integrates with: