Performance Monitoring

Monitor System Performance

Monitoring system performance is one way to help maintain the system. Use vCenter to monitor the following critical Webex CCE components to ensure that the virtual machines perform within system tolerances:

  • CPU

  • Memory

  • Disk

  • Network

Virtual Machine Performance Monitoring

The virtual machines must operate within the specified limits of the Virtual Machine performance counters listed in the following table.

Table 1. Virtual Machine Performance Counters

Category

Counter

Description

Threshold

CPU

CPU Usage (Average)

The CPU usage average in percentage for the VM and for each of the vCPUs.

65%

CPU Usage in MHz (Average)

The CPU usage average in MHz.

95 percentile is less than 65% of the total MHz available on the VM.

Total MHz = vCPUs x (Clock Speed).

CPU Ready

The time a virtual machine or other process waits in the queue in a ready-to-run state before it can be scheduled on a CPU.

150 mSec.

Memory

Memory Usage (Average)

Memory Usage = Active/ Granted * 100

80%

Memory Active (Average)

Memory that the guest OS and its applications actively use or reference. The server starts swap when it exceeds the amount of memory on the host.

95 percentile is less than 80% of the granted memory.

Memory Balloon (Average)

ESXi uses balloon driver to recover memoryfrom less memory-intensive VMs so it can be used by those with larger active sets of memory.

0

Memory Swap used (Average)

ESX Server swap usage. Use the disk for RAMswap.

0

Disk

Disk Usage (Average)

Disk Usage =Disk Read rate + Disk Write rate

Ensure that your SAN is configured to handlethis amount of disk I/O.

Disk Usage Read rate

The rate of reading data from the disk.

Ensure that your SAN is configured to handlethis amount of disk I/O.

Disk Usage Write rate

The rate of writing data to the disk.

Ensure that your SAN is configured to handlethis amount of disk I/O.

Disk Commands Issued

The number of disk commands issued on this disk in the period.

Disk IO per second

IOPS = Disk Commands Issued / 20

Ensure that your SAN is configured to handle this amount of disk I/O.

Stop Disk Command

The number of disk commandsaborted on this disk in the period. The disk command aborts when the disk array takes too long to respond to the command. (Command timeout).

0

Network

Network Usage (Average)

Network Usage = Data receive rate + Data transmit rate

30% of the available network bandwidth.

Network Data Receive Rate

The average rate at which data is received on this Ethernet port.

30% of the available network bandwidth.

Network Data Transmit Rate

The average rate at which data is transmitted on this Ethernet port.

30% of the available network bandwidth.

ESXi Performance Monitoring

The virtual machines must operate within the specified limits of the ESXi performance counters listed in the following table. The counters listed apply to all hosts that contain contact center components.

Table 2. ESXi Performance Counters

Category

Counter

Description

Threshold

CPU

CPU Usage (Average)

CPU Usage Average in percentage for ESXi Server overall and for each of the CPU processors.

60%

CPU Usage in MHz (Average)

CPU Usage Average in MHz for ESXi server overall and for each of the CPU processors.

60% of the available CPU clock cycles.

Memory

Memory Usage (Average)*

Memory Usage = Active / Granted * 100

80%

Memory Used by VMKernel

Memory Used by VMKernel

95 percentile is less than 80% of 2GB.

Memory Balloon (Average)

ESX use balloon driver to recover memory from less memory-intensive VMs so it can be used by those with larger active sets of memory.

0

SwapUsed

ESX Server swap usage. Use the disk for RAM swap.

0

Disk

Disk Commands Issued

Number of disk commands issued on this disk in the period.

Disk IO per second

IOPS = Disk Commands Issued / 20

Disk Commands Aborts

Number of disk commands aborted on this disk in the period.

Disk command aborts when the disk array is taking too long to respond to the command. (Command timeout).

0

Disk Command Latency

The average amount of time taken for a command from the perspective of a Guest OS.

Disk Command Latency = Kernel Command Latency + Physical Device Command Latency.

20 mSec.

Kernel Disk Command Latency

The average time spent in ESX Server VMKernel per command.

Kernel Command Latency should be very small compared to the Physical Device Command Latency, and it should be close to zero.

Network

Network Usage (Average)

Network Usage = Data receive rate + Data transmit rate

30% of the available network bandwidth.

Network Data Receive Rate

The average rate at which data is received on this Ethernet port.

30% of the available network bandwidth.

Network Data Transmit Rate

The average rate at which data is transmitted on this Ethernet port.

30% of the available network bandwidth.

droppedTx

Number of transmitting packets dropped.

0

droppedRx

Number of receiving packets dropped.

0

* The CVP Virtual Machine exceeds the 80% memory usage threshold due to the Java Virtual Machine memory usage.

Monitor Run Time Processing

The following lists the top run time processing recommended for HCS monitoring. Perform these checks to gather the following information:

  • CPU (Average, Peak)

  • Disk Usage (Latency and IOPS)

  • Memory (Average, Peak, Swap, Overhead)

Procedure

Step 1

Review vCenter data to monitor your virtual machine related statistics, such as CPU and memory usage.

This information can be viewed to specific applications that are on the virtual machine. For more information, see the vSphere Monitoring and Performance guide at http://pubs.vmware.com/vsphere-50/topic/com.vmware.ICbase/PDF/vsphere-esxi-vcenter-server-50-monitoring-performance-guide.pdf
Step 2

Review Prime Collaboration Assurance (PCA) data for application and customer specific data for each virtual machine.

Step 3

Review hard disk IOPs to check for spikes, especially during provisioning, backups and similar operations. For more information, see the vSphere Monitoring and Performance guide.

Note 
If you are using a SAN provider, be sure to check the tools provided to check these details. UCS C-series server users should use the VMware tools.
Step 4

Run your Cisco Unified Communications Domain Manager tools to check the health and status of Cisco Unified Communications Domain Manager outside of provisioning times. You can also monitor resource usage for other activities such as Telephony User Interface and End User Self-Care.

Step 5

View capacity reporting data to identify immediate performance concerns and provide trending information based on usage patterns to ensure timely upgrades.

Note 
For more information on capacity, see Cisco Hosted Collaboration Solution Release 12.5 Capacity Planning Guide.

High Availability

It is important to consider High Availability (HA) targets for your system when planning and performing maintenance. The HA targets and maintenance requirements for HCS components may vary from one deployment to another. These targets should be developed as part of the HCS system design process; see the Cisco Hosted Collaboration Solution Release 12.5 Capacity Planning Guide and the Cisco Hosted Collaboration Solution Release 12.5 End-to-End Planning Guide for more information.

In order to meet HA targets, system maintenance is an activity that should occur as much as possible within windows of planned downtime, if downtime is required for the maintenance at all.

Monitor with Prime Collaboration Assurance

Prerequisites

Before you can manage devices using Prime Collaboration Assurance, you need to discover devices in your network. After adding device credentials, you can discover and manage all collaboration endpoints including support for TelePresence components such as Cisco TelePresence Management Suite (TMS). You can also monitor multipoint switches, application managers, call processors, routers, and switches that are part of your network. For more information on Discovering Devices, Managing Device Groups and Managing Inventory see the Cisco Prime Collaboration Assurance Guide - Advanced and the Cisco Prime Collaboration Assurance Addendum for User Guides at http://www.cisco.com/c/en/us/support/cloud-systems-management/prime-collaboration/products-user-guide-list.html.

The table provides a list of network components that Prime Collaboration Assurance can monitor. For more information on Monitoring Devices, see the Cisco Prime Collaboration Assurance Guide - Advanced.

Components Description
Video Endpoints
Endpoint Diagnostics Dashboard Endpoint Diagnostics dashboard displays the details of all video endpoints. You can filter endpoints based on device type, using the Device Group pane.
View User 360 Details

This view displays the end-user information (such as the username, email-id, office phone, and mobile phone numbers) associated with Cisco Unified Communications Manager or TelePresence Management Suite (TMS) endpoints. Photograph and location details for the end user are displayed only if Cisco Prime Collaboration is integrated with LDAP and the username details matches with LDAP details.

Manage a Video Test Call

You can create point-to-point video test calls between two video endpoints in managed state, to test your network. You can see events and alarms, conference statistics, endpoint statistics. Only the CTS, C and EX series codecs are supported for this call.

Monitor Sessions
Data Collection for Video Sessions

Cisco Prime Collaboration periodically polls the following video service infrastructure devices to get information on the sessions: Management devices (CTS-Manager and Cisco TMS), Multipoint Switches, Conference Hardware, and Call and Session Controls (Cisco Unified CM and Cisco VCS).

Import Sessions from CTS-Manager and Cisco TMS

The CTS-Manager and Cisco TMS contain details on the scheduled sessions. Cisco Prime Collaboration periodically polls these devices to retrieve the session details. You can configure the frequency of the periodic polling based on your business needs.

Session Workflow and Scenarios Details on end-to-end scheduled session workflow.

Session Diagnostics Dashboard

Provide details on a session and the endpoints that are involved in that session, troubleshoot and identify faults in the network segment.

You can monitor sessions based on device type, by selecting the desired group from the Group drop-down filter. You can further filter based on the session type using the Show filter in the Video Collaboration Sessions pane.

Monitor Conferences Provides end-to-end details on the conference.
Data Collection for Video Conferences

Cisco Prime Collaboration periodically polls the following video service infrastructure devices to get information on the conferences: Management devices (Cisco TMS), Multipoint Switches, Conference Hardware, and Call and Conference Controls (Cisco Unified CM and Cisco VCS).

Import Conferences from Cisco TMS

The Cisco TMS contain details on the scheduled conferences. Cisco Prime Collaboration periodically polls these devices to retrieve the conference details. You can configure the frequency of the periodic polling based on your business needs.

Conference Workflow and Scenarios

Details on various conference scenarios that are monitored in PCA.

Conference Diagnostics Dashboard

Provides details on a conference and the endpoints that are involved in that conference.

You can monitor conferences based on device type, by selecting the desired group from the Group drop-down filter. You can further filter based on the conference type using the Show filter in the Video Collaboration Conferences pane.

Monitor the Cisco Prime Collaboration Assurance Server

You can monitor the Cisco Prime Collaboration Assurance Server health using the Cisco Prime Collaboration Assurance application. You can get information on CPU, memory, disk utilization, logical storage areas, and process details.