Telemetry

Enable usage monitoring and analytics for agent executions

The --telemetry flag enables detailed usage monitoring and analytics tracking.

Quick Start

praisonai "Your task" --telemetry

Usage

Basic Telemetry

praisonai "Analyze market trends" --telemetry

Expected Output:

📡 Telemetry enabled

╭─ Agent Info ─────────────────────────────────────────────────────────────────╮
│  👤 Agent: DirectAgent                                                       │
│  Role: Assistant                                                             │
╰──────────────────────────────────────────────────────────────────────────────╯

╭────────────────────────────────── Response ──────────────────────────────────╮
│ Based on current market analysis...                                          │
╰──────────────────────────────────────────────────────────────────────────────╯

📊 Telemetry Data:
┌─────────────────────────┬────────────────────────────┐
│ Metric                  │ Value                      │
├─────────────────────────┼────────────────────────────┤
│ Session ID              │ sess_abc123def456          │
│ Start Time              │ 2024-12-16T15:30:00Z       │
│ End Time                │ 2024-12-16T15:30:05Z       │
│ Duration                │ 5.2s                       │
│ Agent Type              │ DirectAgent                │
│ Model                   │ gpt-4o-mini                │
│ Status                  │ success                    │
└─────────────────────────┴────────────────────────────┘

Combine with Metrics

praisonai "Complex analysis" --telemetry --metrics

Expected Output:

📡 Telemetry enabled
📊 Metrics enabled

╭────────────────────────────────── Response ──────────────────────────────────╮
│ [Agent response here]                                                        │
╰──────────────────────────────────────────────────────────────────────────────╯

📊 Combined Analytics:
┌─────────────────────────┬────────────────────────────┐
│ Metric                  │ Value                      │
├─────────────────────────┼────────────────────────────┤
│ Session ID              │ sess_abc123def456          │
│ Duration                │ 8.3s                       │
│ Total Tokens            │ 1,245                      │
│ Estimated Cost          │ $0.0075                    │
│ Model                   │ gpt-4o-mini                │
│ Status                  │ success                    │
│ Tool Calls              │ 2                          │
│ Memory Operations       │ 0                          │
└─────────────────────────┴────────────────────────────┘

Telemetry Data Collected

Category	Data Points
Session	Session ID, timestamps, duration
Agent	Agent type, model used, configuration
Execution	Status, errors, retries
Performance	Response time, token counts
Tools	Tool calls, success/failure rates

Use Cases

Performance Monitoring

Track execution times across different tasks:

# Monitor a complex workflow
praisonai "Multi-step analysis" --telemetry --planning

Debugging

Identify issues in agent execution:

# Verbose telemetry for debugging
praisonai "Failing task" --telemetry -v

Expected Output (with error):

📡 Telemetry enabled

⚠️ Execution Warning:
┌─────────────────────────┬────────────────────────────┐
│ Metric                  │ Value                      │
├─────────────────────────┼────────────────────────────┤
│ Session ID              │ sess_xyz789                │
│ Status                  │ partial_success            │
│ Retries                 │ 2                          │
│ Error Type              │ RateLimitError             │
│ Recovery                │ Automatic retry succeeded  │
└─────────────────────────┴────────────────────────────┘

Usage Analytics

Track patterns over time:

# Run multiple tasks with telemetry
praisonai "Task 1" --telemetry
praisonai "Task 2" --telemetry
praisonai "Task 3" --telemetry

Privacy & Data

Telemetry data is used to improve PraisonAI and is handled according to our privacy policy. No prompt content or sensitive data is collected.

What’s Collected

• ✅ Execution metrics (duration, token counts)
• ✅ Error types and frequencies
• ✅ Feature usage patterns
• ✅ Model selection statistics

What’s NOT Collected

• ❌ Prompt content
• ❌ Response content
• ❌ API keys or credentials
• ❌ Personal information

Disable Telemetry

To disable telemetry globally:

export PRAISON_TELEMETRY=false

Or in Python:

import os
os.environ["PRAISON_TELEMETRY"] = "false"

Best Practices

Development

Enable telemetry during development to catch performance issues early

Production

Use telemetry to monitor production deployments and track SLAs

Debugging

Combine with -v verbose flag for detailed debugging information

Cost Tracking

Pair with --metrics for complete cost and performance visibility

Related

• Metrics CLI
• Telemetry Feature
• Monitoring