Skip to main content

​
Monitoring Dashboard

Transfer Learning includes a real-time monitoring dashboard that provides insights into processing performance, resource usage, and system health.

​
Dashboard Overview

The monitoring dashboard provides a comprehensive view of your Transfer Learning operations:
Transfer Learning Monitoring Dashboard

​
Key Features

Real-time Metrics

View real-time metrics for processing performance, resource usage, and system health.

Historical Data

Access historical data to identify trends and patterns over time.

Alert Management

Configure and manage alerts for critical events and performance thresholds.

Resource Monitoring

Monitor CPU, memory, disk, and network usage during video processing.

​
Accessing the Dashboard

You can access the monitoring dashboard using the following command: 
transfer-learning monitor dashboard
This will start the dashboard server and open it in your default web browser. By default, the dashboard is available at http://localhost:8501.

​
Dashboard Sections

​
Processing Overview

The Processing Overview section provides a high-level view of your Transfer Learning operations:
  • Active Jobs: Currently running processing jobs
  • Completed Jobs: Recently completed jobs
  • Failed Jobs: Jobs that encountered errors
  • Queue Status: Jobs waiting in the processing queue

​
Performance Metrics

The Performance Metrics section displays detailed metrics for video processing:
  • Processing Time: Time taken to process videos
  • Frames Per Second: Processing rate for frame extraction
  • API Response Times: Response times for external API calls
  • Batch Processing Efficiency: Efficiency of batch processing operations

​
Resource Usage

The Resource Usage section shows system resource utilization:
  • CPU Usage: CPU utilization during processing
  • Memory Usage: Memory consumption
  • Disk I/O: Disk read/write operations
  • Network Usage: Network bandwidth utilization

​
Error Tracking

The Error Tracking section provides information about errors and exceptions:
  • Error Rates: Error rates by command and operation
  • Error Types: Distribution of error types
  • Error Timeline: Timeline of errors over time
  • Retry Statistics: Statistics for retry attempts

​
Configuration

You can configure the monitoring dashboard using environment variables or the .env file: 
# Dashboard configuration
DASHBOARD_PORT=8501
DASHBOARD_HOST=0.0.0.0
DASHBOARD_THEME=light
DASHBOARD_REFRESH_INTERVAL=5

​
Integration with External Monitoring

Transfer Learning can integrate with external monitoring systems:
  • Prometheus: Export metrics to Prometheus
  • Grafana: Visualize metrics in Grafana dashboards
  • ELK Stack: Send logs to Elasticsearch for analysis in Kibana
  • DataDog: Send metrics and logs to DataDog