cpu (instance statistics)
The CPU resource lists properties related to overall CPU usage of the AMPS instance. Selecting the items below give more specific information to the type of CPU utilization being consumed by the AMPS process.
Notice that AMPS also collects metrics for the overall host system, available in the host resources.
| Metric | Description | Type |
|---|---|---|
system_percent | Percentage of total host CPU time spent executing kernel code on behalf of this AMPS instance during the statistics interval. | interval average |
total_percent | Sum of user_percent and system_percent for this AMPS instance during the statistics interval. | interval average |
user_percent | Percentage of total host CPU time spent executing user-mode code for this AMPS instance during the statistics interval. | interval average |
AMPS calculates CPU percentages from changes in process and host CPU time counters between samples. Instance percentages use total CPU time across all logical CPUs on the host as their denominator. For example, fully using one logical CPU on a host with four logical CPUs corresponds to approximately 25% of total host CPU capacity.
Values are truncated to whole-number percentages and clamped to the corresponding host CPU percentages to account for sampling differences.
The total_percent metric is calculated when queried through the administrative interface; it is not a stored column in the statistics database. For the SQL expression and source table, see Computed Metrics.
Statistics Database Tables: ICPUS_STATIC, ICPUS_DYNAMIC
Admin Path: /amps/instance/cpu/<metric>