Module 8 – Lesson 4: Monitoring Resource Use Video Transcript
Module 8 – Lesson 4: Monitoring Resource Use Video Transcript
Slide 1
Welcome back! Let’s get started with Lesson 4: Monitoring Resource Use!
Slide 2
After completing this lesson, you should be able to meet the following objectives:
• Monitor the key factors that can affect a virtual machine's performance
• And Use performance charts to view and improve performance.
Slide 3
vCenter Server monitoring tools and guest OS monitoring tools provide different points of
view.
The key to interpreting performance data is to observe the range of data from the perspective of
the guest operating system, the VM, and the host.
The CPU usage statistics in Task Manager, for example, do not give you the complete picture.
View CPU usage for the VM and the host on which the VM is located.
Use the performance charts in the vSphere Client to view this data.
Slide 4
If CPU use is high, check the VM's CPU usage statistics. Use either the overview charts or the
advanced charts to view CPU usage. The slide displays an advanced chart tracking a VM’s
CPU usage.
If a VM’s CPU use remains high over a period of time, the VM is constrained by CPU. Other
VMs on the host might have enough CPU resources to satisfy their needs.
NOT FOR REDISTRIBUTION
Module 8 – Lesson 4: Monitoring Resource Use Video Transcript
If more than one VM is constrained by CPU, the key indicator is CPU ready time. Ready time
refers to the interval when a VM is ready to execute instructions but cannot because it cannot
get scheduled onto a CPU. Several factors affect the amount of ready time:
• Overall CPU use: You are more likely to see ready time when use is high because the
CPU is more likely to be busy when another VM becomes ready to use.
• Number of resource consumers (in this case, guest operating systems): When a host is
running a larger number of VMs, the scheduler is more likely to queue a VM behind
VMs that are already running or queued.
A good ready time value varies from workload to workload. To find a good ready time value
for your workload, collect ready time data over time for each VM. When you have this ready
time data for each VM, estimate how much of the observed response time is ready time. If the
shortfalls in meeting response-time targets for the applications appear largely because of the
ready time, take steps to address the excessive ready time.
Slide 5
To determine whether a VM is being constrained by CPU resources, view CPU usage in the
guest operating system using, for example, Task Manager.
If more than one VM is constrained by CPU, the key indicator is CPU readiness. CPU readiness
is the percent of time that the VM cannot run because it is contending for access to the physical
CPUs.
You are more likely to see readiness values when use is high because the CPU is more likely to
be busy when another VM becomes ready to run. You are also more likely to see readiness
values when a host is running many VMs. In this case, the scheduler is more likely to queue a
VM behind VMs that are already running or queued.
A good readiness value varies from workload to workload.
Slide 6
Compare a VM's memory consumed and granted values to determine whether the VM is
memory-constrained, or not.
Slide 7
If a VM consumes its entire memory allocation, the VM might be memory-constrained, and
you should consider increasing the VM’s memory size.
Slide 8
You might see VMs with high ballooning activity and VMs being swapped in and out by the
VMkernel. This serious situation indicates that the host memory is overcommitted and must be
increased.
Slide 9
Disk performance problems are commonly caused by saturating the underlying physical storage
hardware. You can use the vCenter Server advanced performance charts to measure storage
performance at different levels. These charts provide insight about a VM performance. You can
monitor everything from the VM's datastore to a specific storage path.
NOT FOR REDISTRIBUTION
Module 8 – Lesson 4: Monitoring Resource Use Video Transcript
If you select a host object, you can view throughput and latency for a datastore, a storage
adapter, or a storage path. The storage adapter charts are available only for Fibre Channel
storage. The storage path charts are available for Fibre Channel and iSCSI storage, not for NFS.
If you select a VM object, you can view throughput and latency for the VM’s datastore or
specific virtual disk.
To monitor throughput, view the Read rate and Write rate counters. To monitor latency, view
the Read latency and Write latency counters.
Slide 10
To determine whether your vSphere environment is experiencing disk problems, monitor the
disk latency data counters. Use the advanced performance charts to view these statistics. In
particular, monitor the following counters:
• Kernel command latency: This data counter measures the average amount of time, in
milliseconds, that the VMkernel spends processing each SCSI command. For best
performance, the value should be 0 through 1 millisecond. If the value is greater than 4
milliseconds, the VMs on the ESXi host are trying to send more throughput to the
storage system than the configuration supports.
• And physical device command latency: This data counter measures the average amount
of time, in milliseconds, for the physical device to complete a SCSI command.
Slide 11
Like disk performance problems, network performance problems are commonly caused by
saturating a network link between client and server. Use a tool such as Iometer, or a large file
transfer, to measure the effective bandwidth.
Network performance depends on application workload and network configuration. Dropped
network packets indicate a bottleneck in the network. To determine whether packets are being
dropped, use the advanced performance charts to examine the droppedTx and droppedRx
network counter values of a VM.
In general, the larger the network packets, the faster the network speed. When the packet size is
large, fewer packets are transferred, which reduces the amount of CPU that is required to
process the data. In some instances, large packets can result in high network latency.
When network packets are small, more packets are transferred, but the network speed is slower
because more CPU is required to process the data.
Slide 12
You should now be able to meet the following objectives:
• Monitor the key factors that can affect a virtual machine's performance,
• And Use performance charts to view and improve performance.
This is the end of the Lesson 4 Lecture. If you have any questions, please contact your
instructor. We will see you next time and thanks for watching!
NOT FOR REDISTRIBUTION
