Back to PerformanceManager

Resource Scheduler Counters

Performance counters are modeled using the PerfCounterInfo data object type. Each instance of a PerfCounterInfo data object contains metadata about a specific metric available from the system. The rescpu counter group provides CPU-load-history statistics about resource pools and virtual machines.

VMware vCenter Server stores statistics based on these counters for host’s root resource pool, and for virtual machines configured for high availability (using VMware HA) and VMware DRS (distributed resource scheduler). The VMkernel makes recommendations for relocation of virtual machines based on these statistics. Statistics obtained for a host represent the total of all virtual machines running on the specified host. Statistics include both average and peak values over 1-minute, 5-minute, and 15-minute intervals, as follows:

  • The active counters (actpk1, actpk5, actpk15) summarize the “average” metric values for runtime (runav1, runav5, runav15) and readytime for each virtual machine, during the interval. Statistics based on these counters represent amount of time a virtual machine occupies the CPU during a given interval. The “ready” time is not available as a separate counter, but is tracked and applied by the VMkernel, to the values of the active counters. (“Ready” time is the amount of time during the interval that the virtual machine is ready to run but cannot be scheduled on the CPU for any number of reasons.)
  • The runtime counters (runav1, runav5, runav15, and runpk1, runkpk5, runpk15) represent the average of the actual time the virtual machine is scheduled and running during the interval, and the highest amount of CPU utilization (peak), respectively, during the interval.
  • The maxLimited counters (maxLimited1, maxLimited5, maxLimited15) represent the time that the virtual machine could not run because doing so would violate the limit setting of the virtual machine or resource pool.

   Note:
   The counters on this page are available through the vSphere API only.
   They are not displayed in any of the vSphere Client charts.

Legend:V=VirtualMachine|H=HostSystem|R=ResourcePool|C=ComputeResource|D=OtherDevice
Counter StatsType Unit Level RollupType Label      Entity
VHRC
actav1 absolutepercent3latestCPU Active (1 min. average)οο
CPU active average over 1 minute
actav15 absolute percent 3 latest CPU Active (15 min. average) ο ο
CPU active average over 15 minutes
actav5 absolute percent 3 latest CPU Active (5 min. average) ο ο
CPU active average over 5 minutes
actpk1 absolute percent 3 latest CPU Active (1 min. peak) ο ο
CPU active peak over 1 minute
actpk15 absolute percent 3 latest CPU Active (15 min. peak) ο ο
CPU active peak over 15 minutes
actpk5 absolute percent 3 latest CPU Active (5 min. peak) ο ο
CPU active peak over 5 minutes
maxLimited1 absolute percent 3 latest CPU Throttled (1 min. average) ο ο
Amount of CPU resources over the limit that were refused, average over 1 minute
maxLimited15 absolute percent 3 latest CPU Throttled (15 min. average) ο ο
Amount of CPU resources over the limit that were refused, average over 15 minutes
maxLimited5 absolute percent 3 latest CPU Throttled (5 min. average) ο ο
Amount of CPU resources over the limit that were refused, average over 5 minutes
runav1 absolute percent 3 latest CPU Running (1 min. average) ο ο
CPU running average over 1 minute
runav15 absolute percent 3 latest CPU Running (15 min. average) ο ο
CPU running average over 15 minutes
runav5 absolute percent 3 latest CPU Running (5 min. average) ο ο
CPU running average over 5 minutes
runpk1 absolute percent 3 latest CPU Running (1 min. peak) ο ο
CPU running peak over 1 minute
runpk15 absolute percent 3 latest CPU Running (15 min. peak) ο ο
CPU running peak over 15 minutes
runpk5 absolute percent 3 latest CPU Running (5 min. peak) ο ο
CPU running peak over 5 minutes
sampleCount absolute number 3 latest Group CPU Sample Count ο ο
group CPU sample count
samplePeriod absolute millisecond 3 latest Group CPU Sample Period ο ο
Group CPU sample period.

 

Back to PerformanceManager