Statistics Report

Addax provides detailed statistics and monitoring during job execution to help you understand performance and troubleshoot issues.

Overview

During execution, Addax reports various metrics including:

• Records processed per second
• Bytes transferred per second
• Error counts and percentages
• Channel performance
• Memory usage
• Job progress

Statistics Output

Console Output

During job execution, Addax displays real-time statistics:

2023-12-07 10:30:15.123 [Statistics] INFO - Total records: 15000, Speed: 1500 rec/s (1.2 MB/s), Errors: 0 (0.00%)
2023-12-07 10:30:25.124 [Statistics] INFO - Total records: 30000, Speed: 1500 rec/s (1.2 MB/s), Errors: 2 (0.01%)
2023-12-07 10:30:35.125 [Statistics] INFO - Total records: 45000, Speed: 1500 rec/s (1.2 MB/s), Errors: 2 (0.00%)

Final Report

At job completion, a comprehensive report is displayed:

====================Job Statistics====================
Job ID: 202312071030001
Start Time: 2023-12-07 10:30:00
End Time: 2023-12-07 10:35:30
Total Time: 5m30s

Records:

- Total Read: 100000
- Total Written: 99998
- Errors: 2
- Error Rate: 0.002%

Performance:

- Average Speed: 303 rec/s
- Peak Speed: 450 rec/s
- Total Bytes: 12.5 MB
- Throughput: 38.5 KB/s

Channels:

- Channel 0: 25000 records, 310 rec/s
- Channel 1: 25000 records, 305 rec/s
- Channel 2: 24999 records, 298 rec/s
- # Channel 3: 24999 records, 295 rec/s

Enabling Detailed Statistics

Command Line Option

Enable detailed statistics reporting:

bin/addax.sh -j "-Daddax.stats.detailed=true" job.json

Configuration Setting

Add to job configuration:

{
  "job": {
    "setting": {
      "statistics": {
        "detailed": true,
        "interval": 10000
      }
    }
  }
}

Statistics Components

Performance Metrics

Metric	Description	Unit
Records/sec	Number of records processed per second	rec/s
Bytes/sec	Amount of data transferred per second	B/s
Error Rate	Percentage of failed records	%
Channel Utilization	Performance of each parallel channel	rec/s

Memory Statistics

bin/addax.sh -j "-Daddax.stats.memory=true" job.json

Output includes:

• JVM heap usage
• Memory allocation rates
• Garbage collection statistics

I/O Statistics

Track input/output performance:

bin/addax.sh -j "-Daddax.stats.io=true" job.json

Includes:

• Network I/O rates
• Disk read/write speeds
• Connection pool statistics

Custom Statistics Interval

Set Update Frequency

{
  "job": {
    "setting": {
      "statistics": {
        "interval": 5000
      }
    }
  }
}

Disable Statistics

{
  "job": {
    "setting": {
      "statistics": {
        "enabled": false
      }
    }
  }
}

Export Statistics

JSON Format

Export statistics to JSON file:

bin/addax.sh -j "-Daddax.stats.export=/tmp/job_stats.json" job.json

CSV Format

Export to CSV for analysis:

bin/addax.sh -j "-Daddax.stats.export=/tmp/job_stats.csv -Daddax.stats.format=csv" job.json

Monitoring Integration

Prometheus Metrics

Enable Prometheus endpoint:

bin/addax.sh -j "-Daddax.metrics.prometheus.enabled=true -Daddax.metrics.prometheus.port=9090" job.json

JMX Monitoring

Enable JMX for external monitoring tools:

bin/addax.sh -j "-Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.port=9999" job.json

Performance Analysis

Identifying Bottlenecks

Look for these patterns in statistics:

Reader Bottleneck:

Channel 0: 1000 rec/s (reader working hard) Channel 1: 100 rec/s (waiting for reader) Channel 2: 100 rec/s (waiting for reader)

Writer Bottleneck:

Channel 0: 200 rec/s (waiting for writer) Channel 1: 200 rec/s (waiting for writer)
Channel 2: 200 rec/s (waiting for writer)

Network Bottleneck:

High record rate but low byte rate indicates small records Low record rate but high byte rate indicates large records

Optimization Recommendations

Based on statistics, consider:

1. Low Channel Utilization: Increase channel count
2. High Error Rate: Check data quality and mappings
3. Memory Issues: Adjust JVM heap size
4. Uneven Channel Performance: Check data distribution

Alerting

Error Threshold Alerts

Set up alerts for error rates:

{
  "job": {
    "setting": {
      "errorLimit": {
        "record": 100,
        "percentage": 0.1
      },
      "statistics": {
        "alerts": {
          "errorRate": 0.05
        }
      }
    }
  }
}

Performance Alerts

Alert on low performance:

{
  "job": {
    "setting": {
      "statistics": {
        "alerts": {
          "minSpeed": 100
        }
      }
    }
  }
}

Troubleshooting with Statistics

Common Issues

Job Running Slowly:

• Check channel utilization
• Monitor memory usage
• Verify network connectivity

High Error Rate:

• Review data type mappings
• Check source data quality
• Verify target constraints

Memory Errors:

• Monitor heap usage trends
• Check for memory leaks
• Adjust JVM settings

Debug Commands

Get real-time statistics:

# Monitor job progress
tail -f $ADDAX_HOME/log/addax.log | grep Statistics

# Check JVM status
jstat -gc <pid>

# Monitor network connections
netstat -an | grep <port>

This comprehensive statistics system helps ensure your data synchronization jobs run efficiently and provides the visibility needed for troubleshooting and optimization.