AxonOps Keyspace Dashboard Metrics Mapping

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

Dashboard Overview

The Keyspace dashboard provides keyspace-level performance metrics including latency, throughput, and data distribution. It helps monitor performance characteristics across different keyspaces and identify which keyspaces are consuming the most resources or experiencing performance issues.

Metrics Mapping

Keyspace Performance Metrics

Dashboard Metric	Description	Attributes
cas_Keyspace_ReadLatency	Read operation latency	keyspace, function (percentiles/Count), dc, rack, host_id
cas_Keyspace_RangeLatency	Range query latency	keyspace, function (percentiles/Count), dc, rack, host_id
cas_Keyspace_WriteLatency	Write operation latency	keyspace, function (percentiles/Count), dc, rack, host_id

Data Size Metrics

Dashboard Metric	Description	Attributes
cas_Table_LiveDiskSpaceUsed	Live data size per table	keyspace, scope (table), function=Count, dc

Query Examples

Keyspace Data Size Distribution (Pie Chart)

sum by (keyspace) (cas_Table_LiveDiskSpaceUsed{function='Count',dc=~'$dc'})

Read Latency by Keyspace

cas_Keyspace_ReadLatency{function='$percentile',function!='Min|Max',dc=~'$dc',rack=~'$rack',host_id=~'$host_id',keyspace=~'$keyspace'}

Range Read Latency

cas_Keyspace_RangeLatency{keyspace=~'$keyspace',function='$percentile',function!='Min|Max',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Write Latency

cas_Keyspace_WriteLatency{keyspace=~'$keyspace',function='$percentile',function!='Max|Min',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Read Throughput (ops/sec)

cas_Keyspace_ReadLatency{axonfunction='rate',keyspace=~'$keyspace',function='Count',function!='Min|Max',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Range Read Throughput

cas_Keyspace_RangeLatency{axonfunction='rate',keyspace=~'$keyspace',function='Count',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Write Throughput

cas_Keyspace_WriteLatency{axonfunction='rate',keyspace=~'$keyspace',function='Count',function!='Max|Min',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Panel Organization

Keyspaces Overview Section

• Keyspace Data Size % Distribution - Pie chart showing relative data size per keyspace

Latency Statistics Section

• Read Latency $percentile - Line chart showing read latency at selected percentile

• Range Read Request Latency $percentile - Line chart for range query latency

• Write Latency $percentile - Line chart showing write latency

Throughput Statistics Section

• Reads/sec - Line chart showing read operations per second

• Range Read Requests/sec - Line chart for range queries per second

• Writes/sec - Line chart showing write operations per second

Filters

• data center (dc) - Filter by data center

• rack - Filter by rack

• node (host_id) - Filter by specific node

• groupBy - Dynamic grouping (dc, rack, host_id)

• percentile - Select latency percentile (50th, 75th, 95th, 98th, 99th, 999th)

• keyspace - Filter by specific keyspace(s)

• table (scope) - Filter by specific table(s)

Metric Details

Latency Metrics

• Measured in microseconds
• Available percentiles: 50th, 75th, 95th, 98th, 99th, 999th
• Min and Max values are excluded from queries
• Shows latency at the keyspace aggregation level

Throughput Metrics

• Uses axonfunction='rate' to calculate per-second rates
• Based on the Count function of latency metrics
• Units:
  • Reads: rps (reads per second)
  • Writes: wps (writes per second)

Data Size Distribution

• Aggregates LiveDiskSpaceUsed across all tables in each keyspace
• Shows relative size as percentage in pie chart
• Helps identify keyspaces consuming most storage

Legend Format

• Latency/Throughput panels: $host_id - $keyspace
• Shows both node and keyspace for easy correlation

Performance Insights

Keyspace Comparison:

• Compare latencies across keyspaces
• Identify keyspaces with performance issues
• Monitor throughput distribution

Data Distribution:

• Pie chart shows storage distribution
• Helps with capacity planning
• Identifies data hotspots

Operation Types:

• Separate metrics for reads, writes, and range queries
• Range queries typically have higher latency
• Monitor all three for complete picture

Best Practices

Monitor Latency Percentiles:

• 50th percentile: median performance
• 95th/99th percentile: tail latencies
• Large differences indicate inconsistent performance

Track Throughput Patterns:

• Correlate with application usage
• Identify peak usage times
• Plan capacity accordingly

Data Size Monitoring:

• Regular growth indicates active keyspace
• Sudden changes may indicate issues
• Use for storage capacity planning

Notes

• Keyspace metrics aggregate all tables within the keyspace
• Metrics are collected at the coordinator level
• Range queries include operations like SELECT with ALLOW FILTERING or token ranges