Diamond CPU Metrics

Collected

Fully Qualified Name(FQN) Description Statistic Units Min Max Sparse Data Strategy (SDS) BASE CORR UTIL
cpu.total.guest Percentage of CPU spent running virtual CPUs for guest operatingsystems. average percent 0 none none yes no no
cpu.total.guest_nice Percentage of CPU spent running low-priority virtual CPUs for guestoperating systems. average percent 0 none none yes no no
cpu.total.idle Percentage of CPU not doing any work. average percent 0 none none yes no no
cpu.total.iowait Percentage of CPU time spent waiting for I/O to complete. average percent 0 none none yes no no
cpu.total.irq Percentage of CPU time spent processing hardware interrupts. average percent 0 none none yes no no
cpu.total.nice Percentage of CPU spent running user processes at low priority. average percent 0 none none yes no no
cpu.total.softirq Percentage of CPU time spent processing software interrupts. average percent 0 none none yes no no
cpu.total.steal Percentage of CPU time spent in other operating systems when running ina virtualized environment. average percent 0 none none yes no no
cpu.total.system Percentage of CPU spent running system processes. average percent 0 none none yes no no
cpu.total.user Percentage of CPU spent running user processes. average percent 0 none none yes no no

Computed

Fully Qualified Name(FQN) Description Units Min Max BASE CORR UTIL
metriclyicly.linux.cpu.total.utilization.percent The overall average CPU Utilization across all CPUs. This is anormalized metric. Computation: (data[‘cpu.total.idle’] != null && data[‘cpu.total.idle’].actual!= null) ? ((data.sum(‘cpu.total.*’) – data[‘cpu.total.idle’].actual) /data.sum(‘cpu.total.*’)) * 100 : data[‘cpu.cpu0.idle’] != null ?((data.sum(‘cpu[.].’) – data.sum(‘cpu[.]..idle’)) /data.sum(‘cpu[.].*’)) * 100 : null percent 0 100 yes yes yes
metriclyicly.linux.cpu.total.normalized.user Percentage of CPU spent running user processes. This is a normalizedmetric. Computation: attribute[‘cpus’] == null ? null : data[‘cpu.total.user’] /attribute[‘cpus’].value percent 0 100 yes yes no
metriclyicly.linux.cpu.total.normalized.system Percentage of CPU spent running system processes. This is a normalizedmetric. Computation: attribute[‘cpus’] == null ? null : data[‘cpu.total.system’] /attribute[‘cpus’].value percent 0 100 yes yes no