AxonOps Data Dashboard Metrics Mapping

This document maps the metrics used in the AxonOps Data dashboard.

Dashboard Overview

The Data dashboard provides insights into data storage characteristics including compression ratios, disk space usage, SSTable counts, and partition sizes. It helps monitor data distribution and identify tables with potential issues like large partitions or poor compression.

Metrics Mapping

Compression Metrics

Dashboard Metric	Description	Attributes
cas_Table_CompressionRatio	Compression effectiveness ratio	keyspace, scope (table), dc, rack, host_id
cas_Table_CompressionMetadataOffHeapMemoryUsed	Off-heap memory used by compression metadata	keyspace, scope (table), dc, rack, host_id

Disk Space Metrics

Dashboard Metric	Description	Attributes
cas_Table_LiveDiskSpaceUsed	Live data disk space (excludes deleted data)	keyspace, scope (table), function=Count, dc, rack, host_id
cas_Table_TotalDiskSpaceUsed	Total disk space including tombstones	keyspace, scope (table), function=Count, dc, rack, host_id
cas_Table_LiveSSTableCount	Number of live SSTables	keyspace, scope (table), dc, rack, host_id

Partition Size Metrics

Dashboard Metric	Description	Attributes
cas_Table_MinPartitionSize	Minimum partition size in bytes	keyspace, scope (table), dc, rack, host_id
cas_Table_MeanPartitionSize	Average partition size in bytes	keyspace, scope (table), dc, rack, host_id
cas_Table_MaxPartitionSize	Maximum partition size in bytes	keyspace, scope (table), dc, rack, host_id

Query Examples

Compression Ratio

cas_Table_CompressionRatio{scope=~'$scope', scope!='', dc=~'$dc',rack=~'$rack',host_id=~'$host_id',keyspace=~'$keyspace'}

Compression Metadata Memory

cas_Table_CompressionMetadataOffHeapMemoryUsed{scope=~'$scope', scope!='', dc=~'$dc',rack=~'$rack',host_id=~'$host_id',keyspace=~'$keyspace'}

Live Disk Space Per Table

cas_Table_LiveDiskSpaceUsed{function='Count',dc=~'$dc',rack=~'$rack',host_id=~'$host_id',keyspace=~'$keyspace',scope=~'$scope', scope!=''}

Total Disk Space Per Table

cas_Table_TotalDiskSpaceUsed{function='Count',dc=~'$dc',rack=~'$rack',host_id=~'$host_id',keyspace=~'$keyspace',scope=~'$scope'}

SSTable Count

cas_Table_LiveSSTableCount{dc=~'$dc',rack=~'$rack',host_id=~'$host_id',scope=~'$scope', scope!=''}

Partition Size Metrics

// Minimum
cas_Table_MinPartitionSize{dc=~'$dc',rack=~'$rack',host_id=~'$host_id',scope=~'$scope', scope!=''}

// Mean
cas_Table_MeanPartitionSize{dc=~'$dc',rack=~'$rack',host_id=~'$host_id',scope=~'$scope', scope!=''}

// Maximum
cas_Table_MaxPartitionSize{dc=~'$dc',rack=~'$rack',host_id=~'$host_id',scope=~'$scope', scope!=''}

Panel Organization

Compression Section

• Compression Ratio - Line chart showing compression effectiveness (lower is better)

• Compression Metadata Off-Heap Memory per Table - Memory overhead of compression

Disk Space Per Node Section

• Live Disk Space Per Table - Active data size per table

• Total Disk Space Per Table - Total size including tombstones

• Live SSTable Count Per Table - Number of SSTables per table

Row Size Per Node Section

• Min Partition Size Per Table - Smallest partition in each table

• Mean Partition Size Per Table - Average partition size

• Max Row Size Per Table - Largest partition (identifies potential hotspots)

Filters

• data center (dc) - Filter by data center

• rack - Filter by rack

• node (host_id) - Filter by specific node

• keyspace - Filter by keyspace

• table (scope) - Filter by table

Important Metrics Explained

Compression Ratio

• Shows how well data compresses
• Lower values mean better compression
• Typical values: 0.3-0.5 for text data
• Depends on compression algorithm and data type

Disk Space Metrics

• Live Space: Only counts active data

• Total Space: Includes tombstones and deleted data

• Difference indicates space that can be reclaimed by compaction

SSTable Count

• High counts may indicate:

  • Need for compaction tuning
  • High write load
  • Compaction falling behind

• Affects read performance (more SSTables = more files to check)

Partition Size Distribution

• Min Size: Usually very small (empty or near-empty partitions)

• Mean Size: Average across all partitions

• Max Size: Critical for identifying large partitions

  • Partitions > 100MB can cause performance issues
  • Partitions > 1GB should be investigated

Legend Format

All panels use: $dc - $host_id-$keyspace-$scope

• Shows data center, node, keyspace, and table
• Allows easy identification of specific table metrics

Best Practices

Monitor Compression Ratio:

• Sudden changes may indicate data pattern changes
• Poor compression might suggest wrong algorithm choice

Watch Disk Space Growth:

• Compare live vs total space
• Large differences suggest need for compaction

Track SSTable Counts:

• Consistently high counts impact read performance
• May need to adjust compaction strategy

Monitor Partition Sizes:

• Large partitions (>100MB) need investigation
• Very large partitions (>1GB) can cause operational issues
• Consider data model changes for tables with large partitions

Notes

• The scope!='' filter excludes empty table names
• function='Count' is used for disk space metrics
• All size metrics use binary units (bytes, not SI units)
• Partition size metrics are estimates based on sampling