esxtop Performance Counters

esxtop Performance Counters

Introduction

This article contains a list of some of the performance counters  provided by esxtop. This is far from exhaustive, as this list was  created to answer the question: "which are the most important esxtop  counters?"  Recently VMware has published an exhaustive list of esxtop  information on Interpreting esxtop Statistics

.  Check that out for more information.

CPU Counters

CounterDescription
%RDYThe percentage of time that the world or group is waiting a processor to be available to execute its workload.
%USEDThe percentage of CPU that is used by that world or group.
GIDGroup ID
NWLDThe number of worlds in the group. When this number is greater than one,  the row can be expanded to get information on each world.


Memory Counters

CounterDescription
%ACTVInstantaneous view of the percentage of memory pages that have been used  by the VM in the previous seconds. Unlike TCHD which counts pages by  following working sets, %ACTV is a more frequently updated number that  is based on a sample of the entire memory pool.
%ACTVSSlow moving average of the %ACTV counter.
%ACTVFFast moving average of the %ACTV counter.
MCTL?Set to "Y" when the balloon driver is active in the guest and "N" when not.
MCTLSZThis counter reports the amount of memory that the balloon driver is currently holding for use by other VMs.
MEMSZThe amount of memory (in MB) allocated to the VM at the time of its creation.
NHNThe NUMA home node. This is the node on which the VM is booted.  Migrations that have occurred since the VM started running would result  in this VM running on another node(s).
NMIGThe number of NUMA node migrations since the VM was booted. ESX Server's  scheduler should avoid NUMA migrations so if this number continues to  climb during normal operations some tuning of the VMs may be required.
NRMEMThe amount of memory that exists on a remote NUMA node.
NLMEMThe amount of memory that exists on the local NUMA node.
N%LThe percentage of the VM's memory that exists on the local NUMA node. N%L = NLMEM / (NRMEM+NLMEM)
OVHDThe amount of memory used by the VMkernel to maintain and execute the VM.
SHRDThe amount of the VM's memory that is shared with other VMs.
SHRDSVDThe amount of memory that was saved due to page sharing.  This number  may be less than or equal to SHRD.  As one VM must always claim the  single copy of a shared page, one VM with a shared page will not be able  to claim savings.
SWR/sThe rate at which memory is being swapped in from disk.  High swap rates indicate a need for more memory in the cluster.
SWW/sThe rate at which memory is being swapped out to disk.  High swap rates indicate a need for more memory in the cluster.
TCHDThe amount of memory (in MB) that has been touched (recently used) by  the VM. In this case "recently" means within a minute or two.


Storage Counters

CounterDescription
ABRTS/sThe rate at which disk operations are being aborted. Abort commands are  issued by the guest when the storage system has not responded within an  acceptable amount of time (as defined by the guest OS or application.)
ACTVThe number of IO operations that are currently active. This represents  operations for which the host is processing and can serve as a snapshot  view of storage activity. When this number hovers near zero, the storage  system isn't being used. If is sustains non-zero numbers, the a  constant interaction with the strorage is occurring.
DAVG/cmdThe average amount of time it takes a device (HBA, array, and everything in between) to service a single request.
GAVG/cmdThe total latency seen from the VM when performing an IO operation. GAVG = DAVG+KAVG.
KAVG/cmdThe average amount of time it takes ESX Server's VMkernel to service a  disk operation. Since this number represents time spent by the CPU to  manage IO and processors are orders of magnitude faster than disks, it  should be much, much less DAVG.
QUEDThe number of IO operations that require processing but have not yet be  addressed. Commands are queued and awaiting management by the kernel  when the driver's active buffer is full (see ACTV). Occasionally a queue  will form and result in a small, non-zero QUED number but any  significant (double-digit) average of queued commands means the storage  hardware is unable to keep up with the host's needs.
READS/sThe number of disk reads per second.  READS/s + WRITES/s = IOPS.
WRITES/sThe number of disk writes per second.  READS/s + WRITES/s = IOPS.


Network Counters

CounterDescription
%DRPRXThe percentage of packets that were dropped that was supposed to be received.
%DRPTXThe percentage of packets that were dropped for which transmission was attempted.
MbRX/sThe megabits per second that are received at the network item.
MbTX/sThe megabits per second that are transmitted from the network item.
Comments

Hi Scott,

Thanks for sharing this. I've been searching for the meaning of the following CPU counters:

SWTCH/s

MIG/s

PMIG/s

WMIGI/s

WAKE/s

All are CPU counters in vSphere 4. What values should be considered healthy?

Many thanks from Singapore!

e1

I Agree

Why are those counters not explained? i have been googleing it like crazy without any results.

I got this from support.

SWTCH/s                 switches                number of world switches (out of run state)

The value is typically below 100.

MIG/s                   migrate                 total number of migrations
Migrations can be of two types (a) intra: across the cores on the same socket (b) across cores on different sockets. This captures both. Currently , we dont distinguish between cores on the same socket or across sockets. We distinguish between migrations across cores and migrations across "cells". "cell" is a logical grouping of cores (4 cores by default) done by the cpu scheduler to aid in co-scheduling.

PMIG/s                  processor-migrations    number of inter-core migrations


MIGI/s                  wakeup-migrate-idle     number of migrations on wakeup


QEXP/s                  quantum-expires         number of quantum expirations


WAKE/s                  wakeups                 number of wakeups (from wait state)

Version history
Revision #:
1 of 1
Last update:
‎05-16-2008 11:59 AM
Updated by: