AxonOps System Dashboard Metrics Mapping¶

This document maps the metrics used in the AxonOps System dashboard to their corresponding sources.

Dashboard Overview¶

The System dashboard provides comprehensive monitoring of system resources including CPU, memory, disk, network, and JVM metrics. It helps monitor the health and performance of Cassandra nodes at the operating system level.

Metrics Mapping¶

CPU and Load Metrics¶

Dashboard Metric	Source	Description	Attributes
`host_CPU_Percent_Merge`	System metrics	Overall CPU usage percentage	`time=real`, `dc`, `rack`, `host_id`
`host_CPU`	System metrics	CPU usage by mode	`mode` (system/nice/user/iowait/irq/softirq), `dc`, `rack`, `host_id`
`host_load15`	System metrics	15-minute load average	`dc`, `rack`, `host_id`

Memory Metrics¶

Dashboard Metric	Source	Description	Attributes
`host_Memory_Used`	System metrics	Used memory in bytes	`dc`, `rack`, `host_id`
`host_Memory_UsedPercent`	System metrics	Used memory percentage	`dc`, `rack`, `host_id`
`host_Memory_Cached`	System metrics	Cached memory in bytes	`dc`, `rack`, `host_id`

Disk Metrics¶

Dashboard Metric	Source	Description	Attributes
`host_Disk_UsedPercent`	System metrics	Disk usage percentage	`mountpoint`, `dc`, `rack`, `host_id`
`host_Disk_Used`	System metrics	Used disk space in bytes	`mountpoint`, `dc`, `rack`, `host_id`
`host_Disk_avgqsz`	System metrics	Average queue size	`partition`, `dc`, `rack`, `host_id`
`host_Disk_IOCount`	System metrics	I/O operations count	`partition`, `dc`, `rack`, `host_id`
`host_Disk_SectorsRead`	System metrics	Sectors read	`partition`, `dc`, `rack`, `host_id`
`host_Disk_SectorsWrite`	System metrics	Sectors written	`partition`, `dc`, `rack`, `host_id`
`host_Disk_WeightedIO`	System metrics	Weighted I/O time	`partition`, `dc`, `rack`, `host_id`
`host_Disk_IoTime`	System metrics	Time spent on disk I/O	`partition`, `dc`, `rack`, `host_id`
`host_filefd_allocated`	System metrics	Allocated file descriptors	`dc`, `rack`, `host_id`
`host_filefd_max`	System metrics	Maximum file descriptors	`dc`, `rack`, `host_id`

Network Metrics¶

Dashboard Metric	Source	Description	Attributes
`host_netIOCounters_BytesRecv`	System metrics	Network bytes received	`Interface`, `dc`, `rack`, `host_id`
`host_netIOCounters_BytesSent`	System metrics	Network bytes sent	`Interface`, `dc`, `rack`, `host_id`
`host_ntp_offset_seconds`	System metrics	NTP time offset in seconds	`dc`, `rack`, `host_id`

JVM Metrics¶

Dashboard Metric	Description	Attributes
`jvm_Memory_`	JVM memory usage	`function` (used/max/committed), `scope` (HeapMemoryUsage/NonHeapMemoryUsage)
`jvm_Threading_`	JVM thread information	`function` (ThreadCount)
`jvm_GarbageCollector_*`	GC statistics	`function` (CollectionCount/CollectionTime), collector name

Query Examples¶

CPU Usage per Host¶

host_CPU_Percent_Merge{time='real',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Average IO Wait CPU per Host¶

avg(host_CPU{axonfunction='rate',mode='iowait',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}) by (host_id) * 100

Disk Usage Percentage¶

host_Disk_UsedPercent{mountpoint=~'$mountpoint',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Network Received Bytes Rate¶

host_netIOCounters_BytesRecv{axonfunction='rate',Interface=~'$Interface',Interface!='lo',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

JVM Heap Utilization¶

jvm_Memory_{function='used',scope='HeapMemoryUsage',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

GC Count per Second¶

jvm_GarbageCollector_G1_Young_Generation{axonfunction='rate',function='CollectionCount',dc=~'$dc',rack=~'$rack',host_id=~'$host_id'}

Panel Types and Descriptions¶

CPU usage per host - Line chart showing overall CPU utilization
Avg IO wait CPU per Host - Line chart highlighting I/O wait time
Load Average (15m) - Line chart displaying system load
CPU Usage Detail - Multiple line charts for different CPU modes (User, System, Nice, I/O Wait, Interrupt)
Disk Statistics - Charts for disk usage, IOPS, bytes read/write, queue size
Memory Statistics - Charts for memory usage, JVM heap, GC activity
Network Statistics - Charts for network I/O and NTP offset

Filters¶

data center (dc) - Filter by data center
rack - Filter by rack
node (host_id) - Filter by specific node
mount point (mountpoint) - Filter by disk mount point
partition - Filter by disk partition
interface (Interface) - Filter by network interface

Notes¶

Host metrics are collected by the AxonOps agent at the OS level
JVM metrics are collected via JMX from the Cassandra process
The axonfunction='rate' label calculates rate-based metrics
Network interface 'lo' (loopback) is typically excluded from network metrics
CPU percentages are calculated from rate metrics and multiplied by 100
File descriptor usage is shown as a percentage: (allocated/max)*100