Cisco UCS Performance Manager User Guide

First Published: June 2017

Release 2.5.0

Americas Headquarters
Cisco Systems, Inc.
170 West Tasman Drive
San Jose, CA 95134-1706
USA
http://www.cisco.com
Tel: 408 526-4000
800 553-NETS (6387)
Fax: 408 527-0883
THE SPECIFICATIONS AND INFORMATION REGARDING THE PRODUCTS IN THIS MANUAL ARE SUBJECT TO CHANGE WITHOUT NOTICE. ALL STATEMENTS,
INFORMATION, AND RECOMMENDATIONS IN THIS MANUAL ARE BELIEVED TO BE ACCURATE BUT ARE PRESENTED WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED. USERS MUST TAKE FULL RESPONSIBILITY FOR THEIR APPLICATION OF ANY PRODUCTS.

THE SOFTWARE LICENSE AND LIMITED WARRANTY FOR THE ACCOMPANYING PRODUCT ARE SET FORTH IN THE INFORMATION PACKET THAT SHIPPED WITH
THE PRODUCT AND ARE INCORPORATED HEREIN BY THIS REFERENCE. IF YOU ARE UNABLE TO LOCATE THE SOFTWARE LICENSE OR LIMITED WARRANTY,
CONTACT YOUR CISCO REPRESENTATIVE FOR A COPY.

The Cisco implementation of TCP header compression is an adaptation of a program developed by the University of California, Berkeley (UCB) as part of UCB's public domain version
of the UNIX operating system. All rights reserved. Copyright © 1981, Regents of the University of California.

NOTWITHSTANDING ANY OTHER WARRANTY HEREIN, ALL DOCUMENT FILES AND SOFTWARE OF THESE SUPPLIERS ARE PROVIDED “AS IS" WITH ALL FAULTS.
CISCO AND THE ABOVE-NAMED SUPPLIERS DISCLAIM ALL WARRANTIES, EXPRESSED OR IMPLIED, INCLUDING, WITHOUT LIMITATION, THOSE OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OR ARISING FROM A COURSE OF DEALING, USAGE, OR TRADE PRACTICE.

IN NO EVENT SHALL CISCO OR ITS SUPPLIERS BE LIABLE FOR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, OR INCIDENTAL DAMAGES, INCLUDING, WITHOUT
LIMITATION, LOST PROFITS OR LOSS OR DAMAGE TO DATA ARISING OUT OF THE USE OR INABILITY TO USE THIS MANUAL, EVEN IF CISCO OR ITS SUPPLIERS
HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

Cisco and the Cisco logo are trademarks or registered trademarks of Cisco and/or its affiliates in the U.S. and other countries. To view a list of Cisco trademarks, go to this URL: http://
www.cisco.com/go/trademarks. Third-party trademarks mentioned are the property of their respective owners. The use of the word partner does not imply a partnership
relationship between Cisco and any other company. (1110R)

Any Internet Protocol (IP) addresses used in this document are not intended to be actual addresses. Any examples, command display output, and figures included in the document are shown
for illustrative purposes only. Any use of actual IP addresses in illustrative content is unintentional and coincidental.

© 2014-2017 Cisco Systems, Inc. All rights reserved.

Contents
About this guide.......................................................................................................................4

Chapter 1: Introduction.................................................................................... 5
Dashboard.....................................................................................................................................................7
Events......................................................................................................................................................... 12
Infrastructure.............................................................................................................................................. 14
Reports........................................................................................................................................................28

Chapter 2: Understanding performance and capacity utilization............... 29

Working with service profiles................................................................................................................... 29
Viewing VMware farm storage and compute utilization..........................................................................30
Viewing Hyper-V performance................................................................................................................. 31
Viewing OS performance.......................................................................................................................... 32
Detecting high utilization resources.......................................................................................................... 33
Finding the server with highest utilization................................................................................................34
Determining chassis capacity utilization................................................................................................... 34
Determining blade server capacity utilization...........................................................................................36
Finding congestion between a fabric extender and a rack server............................................................. 37
Exploring domain port channel capacity utilization................................................................................. 39
Determining virtual adapter capacity utilization....................................................................................... 41
Exploring fabric interconnect congestion..................................................................................................41
Exploring fabric utilization and capacity.................................................................................................. 43

Chapter 3: Monitoring network and storage performance and

utilization....................................................................................................... 46
Monitoring and correcting network congestion........................................................................................ 46
Reviewing storage utilization and capacity...............................................................................................49

Chapter 4: Working with capacity projections and finding available

capacity.......................................................................................................... 56
Viewing projected bandwidth exhaustion dates........................................................................................56
Finding remaining capacity for components.............................................................................................57
Finding remaining capacity for an aggregation pool................................................................................ 58
Creating a predictive threshold..................................................................................................................59
Adding a trend line to a graph..................................................................................................................61

3
Cisco UCS Performance Manager User Guide

About this guide

Cisco UCS Performance Manager User Guide provides detailed instructions for performing basic tasks using

■ Cisco UCS Performance Manager Express

■ Cisco UCS Performance Manager

Related publications

Title Description
Cisco UCS Performance Manager Planning Provides general and specific information for preparing to
Guide deploy Cisco UCS Performance Manager.
Cisco UCS Performance Manager Installation Provides detailed information and procedures for installing
Guide Cisco UCS Performance Manager.
Cisco UCS Performance Manager Upgrade Provides detailed procedures for upgrading your existing
Guide Cisco UCS Performance Manager 2.x instance to a newer
version.
Cisco UCS Performance Manager Migration Provides detailed information about, and where applicable,
Guide procedures for migrating data from Cisco UCS Performance
Manager version 1.x into a version 2.x instance.
Cisco UCS Performance Manager Getting Provides instructions for configuring Cisco UCS Performance
Started Guide Manager to monitor your environment after installation.
Cisco UCS Performance Manager User Guide Provides specific instructions for using Cisco UCS
Performance Manager in the UCS environment.
Cisco UCS Performance Manager Provides an overview of Cisco UCS Performance Manager
Administration Guide architecture and features, as well as procedures and examples
to help you use the system.
Cisco UCS Performance Manager Release Describes known issues, fixed issues, and late-breaking
Notes information not already provided in the published
documentation set.

Documentation feedback

To provide technical feedback on this document, or to report an error or omission, please send your comments to
ucs-docfeedback@cisco.com. We appreciate your feedback.

4
 Introduction

Introduction 1
Cisco UCS Performance Manager helps you visualize and monitor your Cisco UCS infrastructure. The product
alerts you to current problems and future potential problems for physical and logical components, network
interconnections, and network performance, including

■ Unified fabric connections between individual components, including upstream dependencies

■ Bandwidth utilization and remaining capacity between various levels of infrastructure hierarchy
■ Physical ports topology
■ Aggregation pools (port channels)
■ Service profiles
■ Organizations
■ IO modules
■ Fabric extenders
■ Chassis
■ Blade and rack servers

The extensive amount of data that Cisco UCS Performance Manager collects and organizes helps you identify
and diagnose different conditions across your infrastructure and enhances your ability to explore potential
solutions. These conditions include

■ Over-subscription of resources
■ Under-performing components
■ Out-of-balance port channel components
■ Connection issues
■ Overload conditions
■ Sources of congestion
■ Component misconfiguration
■ Failing or inappropriately provisioned components

The web-based browser interface provides a wide array of viewpoints, graphs, and reports enable you to
quickly assess the overall health of your entire infrastructure. When necessary, you can drill down to specific
components that require attention. For example, the Dashboard, Topology, and Dynamic View are some of the
views that provide a high-level summary and status of your infrastructure.

Most pages and views contain color-coded indicators and number counts based on system events. These
indicators provide "at-a-glance" information of overall status. The color of the event notification levels indicate
relative severity as follows:

5
Cisco UCS Performance Manager User Guide

■ Red - Critical
■ Orange - Error
■ Yellow - Warning
■ Blue - Info

The following examples show status indicators in the left pane Devices navigation tree and the right pane
Device table.

Figure 1: Devices navigation tree with status indicators

Figure 2: Device table with status indicators

For every component of your infrastructure, whether physical or logical, tabular and graphical information is
available to help you analyze performance and utilization. The information varies by component, but typically
includes utilization and capacity information, and might also include projected exhaustion dates and usage trend
lines.

6
 Introduction

In data tables such as the Device table, you can sort a column from highest to lowest or vice versa. You can add,
remove, and rearrange columns. Using filter fields you can narrow the displayed results; For example, you could
search for device names that begin with "ucs".

Out-of-the-box reports provide current and historical utilization and performance information in tabular and
graphical format. You can refine the content of a report and customize the reported data by using the report
configuration options.

This chapter provides a brief introduction to some of the Cisco UCS Performance Manager views, graphs, and
reports. For more information, refer to the Cisco UCS Performance Manager Administration Guide.

Dashboard
When you log into Cisco UCS Performance Manager, the Dashboard is your initial view. The Dashboard
contains individual portlets that provide a quick view into specific areas of your infrastructure. Initially, it
contains several default portlets, however it is highly customizable. You can add or remove portlets at any time
using the dashboard and portlet controls. You can switch between Dashboards by using the drop-down menu.
Dashboards are specific to individual users, so each user can define one or more customized Dashboards.

Figure 3: Dashboard

The following portlets show information specific to Cisco UCS integrated infrastructure:

■ Domain Overview
■ Fabric Extender Capacity
■ Integrated Infrastructure
■ Service Profile
■ UCS Inventory
■ Welcome to UCS PM

You can see the full list of available portlets when you add a new portlet to the Dashboard.

7
Cisco UCS Performance Manager User Guide

Adding a portlet to the Dashboard

1 In the upper, right corner of the Dashboard, click + > Add Portlet.
The Add Portlet dialog box appears.

Figure 4: Add Portlet

2 In the Add Portlet dialog box, click the drop-down menu and select the portlet.
The Add Portlet dialog box displays the default settings and a preview of the portlet.
3 Modify the portlet configuration options.
4 Click Add.
The portlet is added to the Dashboard.
5 To move the portlet to a different area of the Dashboard, drag and drop it to a new location.

For more information about customizing and creating Dashboards, refer the Cisco UCS Performance
Manager Administration Guide.

Topology view
The Topology view provides a high-level, architectural view of UCS domains and their physical network
connections.

You can launch the Topology view in the following ways:

■ Navigate to the Dashboard and in the upper left corner, click Topology.
■ Navigate to Infrastructure > Devices, in the right pane, select a device, and in the left-side navigation
panel, click Topology.

8
 Introduction

Figure 5: Topology view

The lower left corner shows overall bandwidth and port utilization in the gauges. The view also shows event
information as follows:

■ for all domains (upper left corner)

■ for individual domains that are listed in the Device panel
■ by component

For example, you can view the networked components from fabric interconnects southbound to a chassis, fabric
extenders, and rack servers. Northbound of the fabric interconnects you can see the LAN and SAN clouds.

To work with the Topology view:

■ To see additional information for an item, click the network connection, component, or event notification
icon.
■ To move around the diagram, use the mini-map in the lower, right corner.
■ To zoom in or out, click the +/- controls.
■ To go the event console, in the upper left corner, click the All Domains event bar.

Topology-level status

In the Topology view, status indicators include color-coded network connection lines between components.
For example, a red connection line between two components indicates that a critical event has occurred on that
connection. A colored event icon beside a component indicates that an event has occurred on that component.

9
Cisco UCS Performance Manager User Guide

Figure 6: Topology view: critical event connection

To view additional information about the component or the connection between the components, click an
object or connection. For example, click a connection line between a LAN uplink and the fabric interconnect
component to display information about usage, events, port mappings, and the dependencies of the fabric
interconnect. The projected values show the projected bandwidth usage over the next 30, 60, and 90 days.

10
 Introduction

Figure 7: Ethernet Ports dialog box

Exploring uplink connectivity in Topology view

Using the Topology view, you can explore the uplinks from a fabric interconnect and quickly get detailed
information about the physical connections down to the slot and port numbers. Use the information when you
are troubleshooting poorly performing network connections.

To view network uplinks ports from a fabric interconnect to a LAN cloud:

1 Click Dashboard > Topology.

The topology for the first UCS domain appears.
2 Click a network connection line between a fabric interconnect device and the LAN cloud. (If you have an
Express license, click a different network connection line.)
A window for the connection appears with tabs for Usage, Events, Port Mappings, and Dependencies.
3 Click Port Mappings.
All physical port connections for this network link are displayed.

11
Cisco UCS Performance Manager User Guide

Figure 8:Fabric Interconnect Uplink Port Mappings

In this example, the Port A column shows the fabric interconnect network uplink ports and Port B shows
the LAN switches and ports to which the uplink connects.

These port mappings are discovered by correlating Cisco Discovery Protocol (CDP) information from the
LAN switches with the UCS identity discovered from UCS Manager.

Events
Use the Event Console to view all system-wide events, such as device faults, status events, and performance
threshold events.

Access the Event Console in any of the following ways:

■ In the top-level navigation bar, click Events.

■ From a specific device or component page, in the left pane, click Events.
■ Click a visual event notification, such as an "event rainbow":

Figure 9: Event Console

Note After adding your UCS infrastructure to Cisco UCS Performance Manager, a high number of
threshold-related events might be displayed in the Event Console. These events can be caused by a UCS
configuration that does not align well with the default threshold values for Cisco UCS Performance Manager.

12
 Introduction

You can adjust these thresholds to make your UCS environment easier to view and understand. For more
information about threshold, refer to the Cisco UCS Performance Manager Administration Guide.

Working with the Event Console

The Event Console provides controls that enable you to perform common tasks related to events.

Follow these tips to work with the events table:

■ To view event details, double-click an event row (not the hyperlinks for Resource, Component, or Event
Class).
■ To select multiple events, use Control-click, Shift-click, or Select > All.
■ To change the sort order, click a column heading or enter a value in a filter text box beneath a column
heading:

■ Enter any full string or a subset of a string, optionally with the wildcard (*) contained in the values in that
column.
■ Use "||" (OR), or "!!" (NOT) expressions to further target your filters. For example, type !!status in
the Event Class filter to return all non-status class events.
■ Click the date selection tool in the First Seen and Last Seen columns.
■ Enter a value to match the Count column, as follows:

■ N — Displays events with a count equal to N.

■ :N — Displays events with a count less than or equal to N.
■ M:N — Displays events with a count between M and N (inclusive).
■ M: — Displays events with a count greater than or equal to M.
■ To clear filters, select Configure > Clear filters.
■ To rearrange the column order, drag a column heading to a new location.
■ To manually refresh the events or choose a new refresh cycle, click Refresh.
■ To save the displayed information in CSV or XML format, click Export.
■ To customize the events table, use the Configuremenu options.
■ To perform administrative operations on the selected resource rows, use the Actions and Commands menu
options.

For information about using the Event Archive, Event Classes, and Triggers pages, refer to the Cisco UCS
Performance Manager Administration Guide.

Using the Event Console toolbar

To assist you in managing events in the Event Console, the toolbar provides the following functions:

Figure 10: Event Console toolbar

To perform the following actions, click an event and then click the corresponding toolbar button:

■ Acknowledge an event.
■ Close an event (move it to history).

13
Cisco UCS Performance Manager User Guide

■ Map (associate) an event to a specific event class.

■ Change an event's status to New (revoke its Acknowledged status).
■ Add a log note to an event.
■ Add (create) a simulated event so that you can test a specific condition.

Infrastructure
The Infrastructure page provides controls and menus that help you find, view, and manage devices. The
following figure highlights these tools. To navigate to the Infrastructure page, in the top navigation bar, click
Infrastructure.

Figure 11: Infrastructure page

Review the following terminology that is related to the Infrastructure page:

■ Device — Primary monitoring object in the system. Generally, a device is the combination of hardware and
an operating system.
■ Device class — Special type of organizer that is used to manage how the system models and monitors
devices through the use of monitoring templates.
■ Component — Object contained by a device. Components include interfaces, OS processes, file systems,
CPUs, and hard drives.
■ Integrated infrastructure — A bundle of compute, storage, networking, and virtualization components. Most
integrated infrastructures are bought as one from a vendor:

■ NetApp FlexPod
■ VCE Vblock
■ EMC VSPEX

These infrastructures have UCS as the common compute element, Nexus as networking components, and
VMware as virtualization.

14
 Introduction

■ Managed resource, resource — Servers, networks, virtual machines, and other devices in the IT
environment.

To work with multiple devices, select one or more rows in the table (without clicking a device name or device
class hyperlink). If you click a hyperlink, the corresponding page is displayed. To return to the previous page,
click your browser's back button or Infrastructure.

To work with a single device, click the device's hyperlink name. The device's overview page provides additional
information about the device and access to views, graphs, and the device's individual components.

Exploring the IT infrastructure

The left pane of the Infrastructure page, the Devices panel, provides access to the following device classes and
organizing groups:

■ Devices
■ Host Groups
■ Integrated Infrastructure
■ Component Groups

In the Devices panel, you can drill down in a device class or group to quickly locate a specific device, a group
of common devices or components, or a bundled infrastructure group. To perform the same actions, enter a text
string in the search field at the top of the Devices panel or any of the search fields that are located beneath a
column heading.

Host Groups, Integrated Infrastructure, and Component Groups help you manage your infrastructure in
logical groupings that make sense for your organization.

For more information, see to Working with host groups, Working with integrated infrastructure, and Creating
component groups on page 27

Working with devices and components

From working with a device, you can drill down to work with a component. Use the Overview page and the
Component page.

Figure 12: Overview page

A device's Overview page provides access to different views of the device, Events, an expanded list of
components, graphs, component graphs, custom properties, device administration, and monitoring templates.
The overview pane on the right provides device details; some fields are editable.

15
Cisco UCS Performance Manager User Guide

Figure 13: Component page

The Component page provides multiple panes and graphs. At the top is a list of the components of the type
that you choose. The lower portion of the page displays graphs for the component that you choose in the upper
portion of the page. If graphs are not available for the component, No Graph Data appears.
Perform the following steps to

■ Get detailed information about a specific device.

■ Access different views for a device, such as the Bandwidth Usage, Dependency View, and other related
views.
■ Access the components of a device.
■ Get detailed information about a component.

1 Navigate to the Infrastructure page.

2 On the right, under Device, click a device's hyperlink name.
The devices' overview page appears.
3 In the left pane, select a view, for example Dynamic View.
The right pane is refreshed with the Dynamic View for this device. The left pane continues to display
navigation options for changing the current view, selecting a component, or choosing another option.
4 In the left pane, choose a component, such as logical volumes.
The component page appears. Depending on the chosen component, the page displays a list of components
and graphs.
5 To change the information on the lower half of the page, from the Display drop-down menu, choose an
option, such as Details.

Drilling into device infrastructure

To understand how to drill down into device infrastructure, review the following example. In this example, you
drill down and follow usage from a fabric interconnect back to the server.

16
 Introduction

1 Navigate to the Infrastructure page.

2 In the left pane, choose Devices > CiscoUCS > UCS Manager.
3 In the right pane, click a device hyperlink.
The device's overview page appears.
4 In the left pane, choose Fabric Interconnects.
For the first fabric interconnect in the table, general information and performance graphs are displayed.
5 To see usage and performance information, scroll through the Performance Graphs.
6 From the Display drop-down menu, choose Component Bandwidth Capacity.
The average and maximum Rx and Tx for connected components, such as rack servers, fabric extenders, and
chassis, are displayed.

Maximum utilization is calculated over the last hour. For example, a graph that covers the past six hours
reports maximum utilization for the last hour of that period. Aggregation does not occur.

Figure 14: Component Bandwidth Capacity table

7 To see historical usage, change the Range to Last Hour, Day, or Week.

Figure 15: Range drop-down menu

8 With the component bandwidth capacity table still displayed, to see utilization and capacity graphs, click the
name of a chassis.
Chassis information and the usage and performance graphs appear.

17
Cisco UCS Performance Manager User Guide

Figure 16: Chassis table and graphs

9 To see the busiest servers, click Display > Blade Server Capacity.
The blade servers for the selected chassis appear.
10 To view the blade server performance and utilization graphs, click the name of a blade.
The blade servers table and graphs are displayed.
11 To view service profile information, choose Display > Service Profile Capacity.
The service profile defines a single server and its storage and networking characteristics. Displayed are the
average and maximum utilization for Rx and Tx; the manufacturer and model; number of CPUs, threads, and
ports; operational status; and whether the server is monitored or locked.

Working with graphs

Cisco UCS Performance Manager provides graphs for devices and components. The content of a graph depends
on the selected device or component, and typically contains the data points for the managed resource. When
viewing a component graph, you can show individual data points or all data points simultaneously. For certain
graphs, you can view projected exhaustion dates and trend lines.

To view device and component graphs:

1 Navigate to the Infrastructure page.

2 Choose a device, and in the left pane, click Graphs.
3 To view component graphs, in the left pane, click Component Graphs.

Alternatively, from a device overview page, in the left pane, click Dependency View, and then choose a
table row.

18
 Introduction

Figure 17: Component graph available in the Dependency View

Use the graph tools, labeled in the preceding figure, as follows:

■ To move the graph view forward or back in time, click </>.

■ To narrow or expand the time range that a graph displays, click Zoom In / Zoom Out.
■ To select a predefined time range, click Range.
■ To enter a specific start and end date, choose Custom.
■ To highlight a particular data set, hover over a legend description. You can also click on a legend description
to toggle its display. A solid dot indicates data will be displayed. A hollow dot indicates data will be hidden.
■ To view all data points on the graph at the same time for component graphs, click All on Same Graph.

In the following example, all chassis in a component group appear on the same graph. To see individual
graphs for each component, clear All on Same Graph.

19
Cisco UCS Performance Manager User Guide

Figure 18: Component graph

■ If the graph contains Projections in the lower right corner, hover over the information icon to see projected
exhaustion dates.
■ Choose options from the Action menu (gear) to

■ View the graph definition.

■ Export the graph to a CSV file.
■ Get an HTML link to the graph.
■ Expand the graph in a new standalone window.

For more information about using graphs, see Working with capacity projections and finding available capacity
on page 56 and Adding a trend line to a graph on page 61.

Viewing resource dependencies

The Dependency View shows the resources that are dependent on a device and the resources on which the
device depends.

The following figure shows a sample Dependency View.

20
 Introduction

Figure 19: Dependency View

Accessing the Dependency View

From the Infrastructure page Devices panel, you can access the view in several ways and use the same
functionality.

■ Choose a device to display the device overview page, and then take one of the following actions:

■ Click Dependency View.

■ Expand Components, select a component, and then, from the Display drop-down list, select
Dependencies.
■ Under Host Groups, Integrated Infrastructure, or Component Groups, click an object. At the top of the
pane, click Details, and then click Dependency View.

Working with the Dependency View

■ To expand or collapse the contents of a group, click + or -.

■ To save a copy of the information, click Export.
■ To toggle the view between resources that depend on the component to resources on which the device
depends, click Dependents or Dependencies.

If a resource has no dependents, a message appears.

■ To show components that are greater than or equal to a certain utilization percentage, use the Utilization
Filter. The default setting is 0%, which shows all dependents. Moving the slider to the right increases
the utilization percentage. As you move the slider, a tooltip displays the current percentage and the list of
dependents changes.

Components that have open events are automatically displayed regardless of theUtilization Filter setting.
■ To change the list of displayed groups, click Groups. Unchecked groups are not displayed. To add a group
back to the table, check the group name.

Viewing infrastructure hierarchy and resource relationships

The Dynamic View displays an architectural overview of your infrastructure and relationships to other
resources. Event information for each object is shown in the infrastructure hierarchy.

The Dynamic View is available for

■ Server devices (registered OS)

■ Integrated infrastructure
■ Host groups, which include a dynamic view for vSphere

21
Cisco UCS Performance Manager User Guide

To access the Dynamic View:

1 Navigate to the Infrastructure page.

2 Select a device, integrated infrastructure, or host group.
3 In the left pane, click Dynamic View.

The following figure shows the Dynamic View.

Figure 20: Dynamic View

4 To change the perspective, save the image, or refresh the image, use the controls in the upper right corner.

Figure 21: Dynamic View controls

5 To access the Inspector, click an object.

The Inspector displays object events, device status, production station, and location.

Figure 22: Inspector

22
 Introduction

Viewing bandwidth usage and projections

The Bandwidth Usage View shows usage information and projected exhaustion date ranges for the following
metrics:

■ Average Rx, Average Tx

■ Maximum Rx, Maximum Tx
■ Average LAN Tx, Maximum LAN Tx
■ Average Storage Tx, Maximum Storage Tx
■ Projected Tx Date, Projected Rx Date (The number of days before the maximum capacity will be reached.)

To access and work with the Bandwidth Usage View:

1 Navigate to the Infrastructure page.

2 Select a device to display the device's overview page.
3 In the left pane, click Bandwidth Usage.

Figure 23: Bandwidth Usage View with graphs expanded

Figure 24: Bandwidth Usage View with graphs collapsed

To work with the Bandwidth Usage View:

■ To see server devices and components, click Server.

■ To see network devices and components, click Network.
■ To drill down into lower-level infrastructure, click the expand/collapse arrows to the right of the resource
name.

23
Cisco UCS Performance Manager User Guide

Note The aggregation pool values (such as the top level chassis/fex) are calculated at an interval of 10
minutes, which might result in some variance compared to the sum of the values of the underlying parts.
■ To view additional columns, slide the horizontal scroll bar to the right.
■ To hide or show graphs, click the collapse/expand button in the upper right corner of the graph.
■ To customize the column headings in the table, click the down arrow to the right of any column heading, and
then select Columns.
■ To select the past hour, past six hours, or the past day, click the Time Range drop-down menu.

To work with graphs:

■ To highlight information in a graph, hover over a data source at the top of the graph.
■ To toggle information on and off in a graph, click a data source at the top of the graph.
■ To see information for a specific point in time, hover over the graph.
■ For a larger view, pop out the graph in a new tab.
■ To change the graph's time line, click Zoom in, Zoom Out, <, and >.

Organizing your monitored infrastructure

IT administrators are typically responsible for the administration and management of hundreds, or even
thousands, of monitored devices. In addition, each monitored device might contain hundreds of individual
components, which must also be monitored. Tracking events and overall status for every resource can quickly
become overwhelming.

To ease the task of monitoring large numbers of devices and components, Cisco UCS Performance Manager
provides a holistic view into the health and status of monitored devices by using the following organizational
containers. These containers (groups) are easy to create, organize, and traverse, and can be based on company
organization, function, or location.

■ Host groups
■ Integrated infrastructure groups
■ Component groups
Host groups enable you to create a logical grouping of Windows and Linux servers or VMs. Using host groups,
you can easily see events for the group of devices as a whole. You can set up host groups by

■ Operating system type; for example, create a group for all Windows devices and another group for all Linux
devices.
■ Geographic location; for example, create a group for all devices in Austin and another group for all devices
in London.
■ Function; for example, create separate groups for devices in the Finance department, the Marketing
department, Production systems, Test systems, and so on.

Integrated infrastructure groups provide a convenient way to view the system as a combined resource instead
of as individual pieces. Integrated infrastructure is a collection of the following device types that function as an
optimized computing package.

■ Compute, which includes Cisco UCS and UCS Mini devices that are the primary source of servers or VMs.

Servers and VMs that run operating systems are not considered compute resources, and therefore cannot be
added to an integrated infrastructure group.
■ Network, including supported switches such as the Cisco Nexus series.
■ Storage, including storage devices such as the EMC and NetApp devices.
■ Virtualization, including Hyper-V or VMware hypervisors.

24
 Introduction

Note The integrated infrastructure functionality is only available with a Cisco UCS Performance Manager
license.

Component groups provide a logical way to view and manage individual components of a monitored device.
For example, a Cisco UCS device can have many components such as Ethernet ports, blade servers, aggregation
pools, and so on. When you add components to a component group, you can view all events for the group in a
single location; display component graphs with all components on the same graph; view the group's dependents
and dependencies; monitor or lock individual components in the group or the entire group. Components that you
add to a group maintain their original permissions.

The following example shows a component group that consists of the chassis group, with two nested groups for
specific geographical locations.

Figure 25: Component groups

Creating host groups

Create a host group of devices that you are already monitoring.

1 Navigate to the Infrastructure page.

2 In the left pane, click Host Groups.
3 In the lower-left portion of the left pane, click Add.
The Add Group dialog box appears.
4 Enter a name and description of the host group (for example, Production or Austin), and then click Submit.
The new name appears under the Host Groups device class. To refresh the tree, right-click the device class
and choose Refresh Tree.
5 To display a list of monitored devices, click Devices or a device class.
6 Click the row of the server (OS) devices that you want to add to the host group.

Note Be sure to click anywhere on the row that is not a hyperlink to select it. If you click a hyperlink,
you will be taken to the specific details page.
7 Drag-and-drop the selected devices on the name of the host group and accept the move action.

Creating integrated infrastructure groups

Use an integrated infrastructure group to understand and detect issues within that infrastructure.

Note You cannot add individual servers or VMs to an integrated infrastructure group; however, you can add
them to host groups. For more information, see Working with host groups.

25
Cisco UCS Performance Manager User Guide

1 Navigate to the Infrastructure page.

2 In the left pane, click Integrated Infrastructure.
3 With the integrated infrastructure device class selected, at the bottom of the Devices panel, click Add.
The Add Integrated Infrastructure dialog box appears.
4 Enter a name and description, and then click Submit.
The new group appears under the Integrated Infrastructure device class. To refresh the tree, right-click the
device class and choose Refresh Tree.
5 Add members to the new integrated infrastructure as follows:
a In the left pane, click Devices.
b Select the rows of the devices that you want you to add, and then drag and drop the rows on the new
integrated infrastructure group.
c Click OK.
The group displays devices as they are added, including a list of open events and their severity, as shown
in the following figure.

Figure 26: Example Integrated Infrastructure Group

6 After you create and populate an integrated infrastructure group, you can display information about
the group as a bundled set of resources. For examples, perform the following steps on an integrated
infrastructure group;
a In the left pane, click Dynamic View.
The Dynamic View displays the hierarchy of the components, organized by type, and components that
have open events. (See
b Under Integrated Infrastructure, highlight the integrated infrastructure name, and then click Details.
The Devices pane displays quick-access links as follows:

■ Devices
■ Events
■ Device Administration
■ Dependency View
■ Dynamic View

Using the Integrated Infrastructure portlet

For a quick summary of the overall health of the major components of your integrated infrastructure, use the
Integrated Infrastructure portlet. The porlet summarizes event information for each area of your integrated
infrastructure: storage, network, virtualization, and compute.

26
 Introduction

Figure 27: Integrated Infrastructure portlet

To use the portlet, complete the following tasks.

1 Create an integrated infrastructure group to populate the portlet, as described in Creating integrated
infrastructure groups on page 25.
2 Add the Integrated Infrastructure portlet to the dashboard, as described in Adding a portlet to the
Dashboard on page 8.
3 From the Integrated Infrastructure drop-down menu, choose an integrated infrastructure or host group.
4 Navigate to the Event Console.
5 Click a resource or component to go the overview page.
The overview page provides access to additional information, including graphs, component graphs,
dependencies, and other components. The available options vary based on the resource type.
6 Continue to drill down in any area of interest until you reach the cause of the reported event.

Creating component groups

1 Navigate to the Infrastructure page.

2 In the left pane, click Component Groups.
3 In the lower-left portion of the left pane, click Add.
The Add Component Group dialog box appears.
4 Enter a name and description, and then click Submit.
The component group appears under the Component Groups device class. To refresh the tree, right-click
the device class and choose Refresh Tree.
5 In the left pane, select the component group name.
6 In the right pane under the component group name, click Add.
The Add to Component Group dialog box appears.
7 In the search field, enter a component type, such as blade servers.
The Search Results area displays matching components.
8 Click the row (not the hyperlink name) of one or more components, and then click Add.
9 To save the group and close the dialog box, click Close.
10 To view component graphs:
a In the left pane, click the group name.
b At the top of the left pane, click Details.
c Click Component Graphs, and then check All on Same Graph.
For more information, see Working with graphs on page 18.
11 To view events for the component group, click Events.
12 To disable monitoring of one or more components:
a In the upper right corner of the left pane, click See All.
b Select the components to disable.
c Click Action (gear) and choose Monitoring.
d In the Monitoring pop-up window, click Yes.

27
Cisco UCS Performance Manager User Guide

Reports
Reports provide historical performance data in a tabular or graphical format. The following report categories
are available. (If you are using a Cisco UCS Performance Manager Express license, a smaller set of reports is
available.)

■ Cisco UCS Capacity Reports

■ Cisco UCS Reports
■ Enterprise Reports
■ Monitoring Capabilities Reports
■ Performance Reports
■ System Reports
■ vSphere

Each report category contains multiple reports. Each report has several options that help you define the content
of the report. Procedures presented later in this guide explain how to create reports. For more information, refer
to the Cisco UCS Performance Manager Administration Guide.

Generating a report
Cisco UCS Performance Manager reports are grouped into categories. Use the report parameters to generate a
report with a specific set of data.

1 In the top navigation bar, click Reports.

2 Expand a report category and select a specific report.
The report is generated using the default parameters.
3 To see the reports, scroll down.
4 To customize the displayed data, configure the report options.
5 To export the report to a CSV file, click export.

28
 Understanding performance and capacity utilization

Understanding performance and capacity

utilization 2
This chapter shows you how to use Cisco UCS Performance Manager to analyze integrated infrastructure
performance. Example procedures illustrate how to determine whether resources can handle additional workload
using remaining capacity, are close to capacity, or have reached full capacity.

Cisco UCS Performance Manager enables you to analyze system resources and components to identify potential
versus actual capacities. For example, you can

■ Determine how close a component is to maximum capacity, and plan for potential expansion or restructure.
■ Determine a component's amount of available remaining capacity, and whether a device is over-subscribed
or has additional resources that you can use to alleviate over-subscription or help eliminate bottlenecks.
■ Determine whether IO module server ports, Ethernet uplinks, and FC uplinks are or have been congested,
and act to alleviate the congestion or potentially forecast the next cyclic congested event.
■ Identify sources of and explore ways to address congestion, such as moving service profiles between chassis.
■ Determine whether virtual or physical workloads and applications are affected by UCS server CPU and
memory configuration, and plan for or make changes to server configurations or hardware components.
■ Compare current and historical performance across converged infrastructure components to identify where
constraints exist and decide whether and how to remove them.

Working with service profiles

A service profile is a software definition of a server and its LAN and SAN network connectivity. It defines a
single server and its storage and networking characteristics. The Service Profile portlet provides an at-a-glance
status of all service profiles that are modeled in a specific domain or across all domains. To access detailed
information for a specific profile, in the portlet, click a profile name.

Figure 28: Service Profile portlet

29
Cisco UCS Performance Manager User Guide

Filter and sort features enable you to work efficiently with many profiles. Filtering works in real-time as you
type characters in the text field. Filter on any columns that contain a text string, such as Name, Server, Domain,
Type, including hidden columns. Sorting works for all columns.

For each profile, the Service Profile portlet contains the following information:

■ Events — Displays the highest-level severity event. For detailed event information, navigate to the Event
Console.
■ Name — Displays the service profile name.
■ Operational state — Displays the operational state of the server as defined in the service profile.
■ Domain — Displays the UCS domain name.
■ Server — The blade or rack server as defined in the service profile.
■ Logical server — The related device associated with the service profile. Related devices can be hypervisors
(Hyper-V or vSphere), or a Windows or Linux server.
■ Type — Indicates the type of logical server as defined in the service profile, as follows:

■ Hyper-V
■ vSphere
■ Windows server
■ Linux server
■ Organization — Displays the container hierarchy, for example, organization or root/tier/
server name.
■ Local — A flag indicating whether a service profile comes from UCS central.

By default, the Service Profile portlet shows average utilization for

■ CPU
■ LAN Tx
■ LAN Rx
■ SAN Tx
■ SAN Rx

For hypervisors, you can also display hidden columns showing maximum utilization for CPU, LAN, and SAN.
To view statistics for Windows and Linux servers, click the link in the Logical Server column.

1 To add the Service Profile portlet to the dashboard, follow the steps in Adding a portlet to the Dashboard on
page 8.
2 To change the default view:
a To view a different domain, select it from the Domain drop-down menu.
b To view all domains, select Domain > All Domains.
c To find a specific service profile in a domain, in the Filter text field, enter the first few characters of the
name.
d To change the sort order, click any column heading and then click Sort Ascending or Sort Descending.

Viewing VMware farm storage and compute utilization

Use Cisco UCS Performance Manager to identify oversubscribed resources and those that have extra capacity.
Without using vSphere or providing vSphere credentials, you can view the top datastores by percentage used,
and the top hosts based on CPU and memory usage.

30
 Understanding performance and capacity utilization

Figure 29: VM storage and compute statistics

To view VMware farm information:

1 Navigate to the Infrastructure page.

2 In the left pane, choose Devices > vSphere.
3 Click a vSphere device name.
4 For OS information, scroll down and on the right, review the OS Model information.

Viewing Hyper-V performance

For Hyper-V devices, you can view performance graphs and dependency information, and drill down to detailed
component information.

1 Choose Infrastructure > Devices > Server.

The server list appears on right side of the page.
2 Select a Hyper-V device.
If necessary, expand the Server > Microsoft category or enter the string Hyper-V in the search box
beneath the Device Class column heading.
The overview page for the Hyper-V device appears.
3 To view performance information, in the left pane, click Graphs.

31
Cisco UCS Performance Manager User Guide

Figure 30: Hyper-V performance graphs

4 To see additional views or more information, use navigation tools in the left pane.

Viewing OS performance
Cisco UCS Performance Manager enables you to survey and monitor operating system performance. Review
historical and real-time performance and usage data of components such as CPU and memory.

To review OS performance information:

1 Choose Infrastructure > Devices > Server.

2 Drill down the server categories to find a server, such as Linux server, and click the device name.
The device overview page appears, and describes the OS model and version.
3 In the left pane, click Graphs.

The following graphs are displayed:

■ Load Average
■ CPU Utilization
■ Memory Utilization
■ I/O

32
 Understanding performance and capacity utilization

Figure 31: Operating system performance graphs

4 To customize the view, use the graph tools and Range drop-down menu.

Detecting high utilization resources

To quickly locate resources with high utilization, use the Dependency View. The Utilization Filter helps you
focus on resources that meet a specified utilization percentage. The filter setting is saved per user, per device,
until you change it.

Figure 32: Dependents with >=80% utilization

1 Navigate to Infrastructure > Devices, and expand Cisco UCS.

2 Select a UCS Manager device.
3 In the device's overview page, click Dependency View.
Dependents of the device are displayed. By default, all resources that are currently using >=0% utilization
are displayed.
4 Slide the Utilization Filter to the right until it reaches >=80%.
Resources that match the filter are displayed at the top of the table, followed by all resources that have a
warning, critical, or error status, regardless of their current utilization.
5 To see resources that depend on the device, click Dependents.
The utilization filter remains at >=80%.
6 View and work with graphs for a component. For example:
a To expand or collapse the components of a group, click + beside the group (such as Ethernet Ports).
Alternatively, click Groups and select a group.

33
Cisco UCS Performance Manager User Guide

b Click on an Ethernet port row (not the component name link).

If graphs are available for the selected resource, they are displayed on the right.
c To analyze the graph data, use the graph tools.
d To save the information to a CSV file, click Export.

Finding the server with highest utilization

To find the server in a domain with the highest bandwidth usage, see the Dashboard. The following portlets
include utilization information:

■ Domain Overview — Shows utilization and open event counts.

■ Service Profiles — Shows current fabric utilization across all domains or any single domain.
■ Chassis Capacity — Shows utilization by chassis for a single domain.
■ UCS Inventory — Shows total number of blades and how many are associated with service profiles.

Determining chassis capacity utilization

To determine whether a selected chassis is used to its full capacity and to explore chassis information, use the
Topology view.

1 Navigate to Dashboard > Topology.

2 In the left pane, click a UCS domain.
3 In the Topology view, double-click a fabric interconnect to chassis connection.
A popup window for the switch to chassis server ports appears, as shown in the following example.

34
 Understanding performance and capacity utilization

Figure 33: Ethernet Pool Bandwidth Capacity Utilization graph

The Usage tab graphs include values for capacity utilization, projected high utilization percentage, projection
dates, and total data transferred in Gbps. In the example graph, the Avg Rx value is 24.5%, which leaves
bandwidth for additional transactions.

The Ports area shows that four ports connect from Switch-A to Chassis-3. The Bandwidth column shows
that each port has a maximum capacity of 10 Gbps. Therefore, total capacity is 40 Gbps for both receiving
and sending data.
4 Review and work with Usage tab information and graphs as follows:
a To display the threshold value, click Projected High Pct Rx.
b Review the 30, 60, and 90 day Projected Values.
Projections are calculated on data that was collected over the last 10 days. A value of N/A or zero means
that data is insufficient to make a projection.
c Scroll to the Ethernet Pool Utilization graph that shows total throughput.
In the following example, the total average of received and sent data is 19.6 Gbps out of a total capacity
of 40 Gbps.

35
Cisco UCS Performance Manager User Guide

Figure 34: Ethernet Pool Utilization graph

Determining blade server capacity utilization

The Chassis page displays graphs that identify which blade servers are the heaviest users of chassis bandwidth
within the chassis.

1 Navigate to Infrastructure > Devices.

2 In the right pane, click the name of a device.
3 In the left pane, click Components > Chassis.
4 In the left pane, click Bandwidth Usage.
On the right, the Bandwidth Usage page appears.
5 To list the components of a chassis, click the expansion arrow.
In the following example, the projected columns show when the chassis will reach 64 Gbps based on the
average Tx utilization over the last ten days.

36
 Understanding performance and capacity utilization

Figure 35: Bandwidth Usage for a chassis

6 Review the columns for total bandwidth, averages for received and sent data, and the Projected Tx Date and
Projected Rx Date columns.

Note The aggregation pool values (the top level chassis or FEX) are calculated at an interval of ten
minutes, which might result in variance compared to the sum of the values of the underlying parts.
7 Click the row of a component (not the hyperlink name), and review the available graphs; for example, Blade
Server Utilization and Service Profile Utilization .

Finding congestion between a fabric extender and a rack server

To determine whether network congestion exists between a fabric extender and a rack server, use the Topology
view and analyze the throughput, capacity, and remaining capacity values.

1 Navigate to Dashboard > Topology.

2 On the left in the All Domains pane, click a UCS device.
The device's topology appears.
3 Double-click a connection line between a fabric extender and a rack server.
A fabric extender to rack server window appears.

In the following example, the Ports sections shows that the connecting port's capacity is 10 Gbps.

37
Cisco UCS Performance Manager User Guide

Figure 36: Fabric extender to rack server usage example

4 Scroll to and review the Ethernet Pool Utilization graph.

The following example shows that minimal data is being sent and received, leaving almost the full capacity
of the 10 Gbps port available for additional transactions. In this example, congestion is not an issue.

38
 Understanding performance and capacity utilization

Figure 37: Ethernet Pool Utilization graph example

Exploring domain port channel capacity utilization

You can access current and historical data about capacity usage of domain port channels and drill down into
each component within a domain.

1 Navigate to Infrastructure > Devices > CiscoUCS.

2 Click a device name.
3 To view LAN uplink information, in the left pane, click Components > Aggregation Pools.

39
Cisco UCS Performance Manager User Guide

Figure 38: Ethernet pool bandwidth capacity utilization graph

4 To view information about the uplink members, click Display > Members.
The aggregation pool members are displayed.
5 From the aggregation pools table, click a chassis name.
The ports for each switch on the chassis are displayed.

Figure 39: Chassis port members

6 To view graphs, click Display > Graphs.

40
 Understanding performance and capacity utilization

Determining virtual adapter capacity utilization

View information about the virtual adapters in your infrastructure, including the associated service profiles and
historical capacity data.

1 Navigate to Infrastructure > Devices > CiscoUCS.

2 Select a Cisco UCS device.
3 Click Components > > Service Profiles.
4 In the right pane, select a profile.
5 Click Display > vHBA Capacity.
The table of vHBAs appears and shows the average and maximum utilization for Rx and Tx.
6 To see more information about an adaptor, click the adaptor name.
The performance graph for the vHBA appears.

Figure 40: Virtual HBA performance graph

7 To view historical data, on the performance graph toolbar, click Range.

8 To view virtual NIC usage, in the left pane, click Components > Virtual NICs.
9 To view all NICs for a specific profile, sort the Service Profile column.

Exploring fabric interconnect congestion

The Topology view shows a graphical interpretation of a switch performance and tabular data for each Ethernet
port. The tabular data includes the port slot location, bandwidth, and utilization information. If additional
information is required, you can drill down to specific components. The following example shows how to select
a network component and drill down into the performance data for its Ethernet ports.

1 Navigate to Dashboard > Topology.

The Topology view appears.
2 Click a component, such as a fabric interconnect.
The following example shows fabric interconnect usage information.

41
Cisco UCS Performance Manager User Guide

Figure 41: Fabric interconnect usage information

3 Analyze the data to determine whether the component and its associated connections are performing
appropriately.
Identifying and correcting components that do not perform as required helps to prevent and alleviate
congestion in the network. The following example shows the relative congestion of the LAN uplink to
switch-A connection. Because the last values are 3.27 Mbps for received and 2.02 Mbps for sent data, they
are not yet close to the maximum available bandwidth of 10Gbps, so congestion is very low. If these values
increase to approach the 10Gbps maximum available bandwidth, congestion increases and eventually leads
to network issues.

42
 Understanding performance and capacity utilization

Figure 42: Congested Switch A Uplink Ethernet Ports

4 To display the overview page for a component, click the component name.

Exploring fabric utilization and capacity

You can view and analyze component and port utilization information for a fabric interconnect. The Fabric
Interconnects page displays graphs that show health checks, bandwidth utilization, remaining capacity,
northbound utilization, LAN Cloud utilization, SAN Cloud utilization, and direct storage utilization. Graphs
help you identify the heaviest consumers of bandwidth in the domain. Remaining capacity data indicates how
much unused bandwidth capacity is available for each chassis.

1 Navigate to Infrastructure > Devices.

2 Select a Cisco UCS device.
The device's overview page appears.
3 Click Components > Fabric Interconnects.
In the right pane, the Fabric Interconnects page appears.
4 Select a fabric interconnect row, and then click Display > Graphs.
5 Scroll down to the Fabric Interconnect Utilization (bits) graph.
In the following example, received bits per second is 40.79 Mbps on average, and sent bits per second is
approximately 17.5 Mbps.

43
Cisco UCS Performance Manager User Guide

Figure 43: Fabric Interconnect Utilization (bits) graph

6 Scroll down to the Fabric Interconnect Remaining Capacity graph.

In the following example, average utilization is 390.96 Gbps, which leaves approximately 50% remaining
capacity.

Figure 44: Fabric Interconnect Remaining Capacity graph

7 Scroll through the graphs to see utilization by direction to the LAN Cloud, SAN Cloud, direct-attached
storage, and northbound utilization.
8 To display data for the components or ports within the fabric interconnect, click Display, and then click
Component Bandwidth Capacity or Ports Capacity.
In the following example, the table displays average sent and received utilization percentage for each
chassis.

44
 Understanding performance and capacity utilization

9 In the data tables, you can review and work with data; for example:
a Sort a column from highest to lowest utilization or vice versa.
b To search for a role, such as server, in the Type to filter field, enter the role.
c To display a different time/date range, click Range and select Last Hour, Last Day, or the Last Week.

45
Cisco UCS Performance Manager User Guide

Monitoring network and storage

performance and utilization 3
This chapter contains concepts and procedures to help you understand current and historical utilization and
capacity of your network and storage infrastructure.

Monitoring and correcting network congestion

Network congestion occurs when an increase in data transmissions results in a proportionately smaller increase
or throughput reduction across the network. Congestion can result from more data being sent across a network
than it can handle.

Applications send data as packets across the network, traversing network devices such as fabric extenders, fabric
interconnects, Ethernet ports, routers, and switches. The buffers on over-subscribed devices can fill up and
overflow, and data packets are lost. This state can cascade into even greater congestion because applications
must retransmit lost packets, which results in additional data traffic across the network.

If this cyclic cascade continues, the network can become paralyzed and fall into a state of congestive collapse.
This state that ensues when congestion increases to the point that throughput drops to and remains at very low
levels, and adversely affects the performance of applications and devices.

Correcting or preventing congestion is important to keeping a network running smoothly for users and
applications. Use Cisco UCS Performance Manager tools to analyze collected data to identify active network
congestion issues and make informed decisions about how to correct the issues.

The collected data includes projected, current, and historical performance data, and component, connection,
and provisioning data. By analyzing that data, you can identify and isolate for rehabilitation any potential
inadequacies or pinch points in your network, and plan for and prevent future network congestion issues.

Use Cisco UCS Performance Manager tools to

■ Identify and reconfigure service profiles that contribute to low performance.

■ Move service profiles between servers to enhance server performance.
■ Identify and rectify server provisioning, including processor and memory configurations.
■ Locate and correct bandwidth inequities. For example, identify network components with critically small
available remaining capacity. To bring the network in balance, redistribute bandwidth between components
with appropriate available remaining capacity or add equipment.
■ Use historical data from the various performance logs to determine patterns and cycles to plan for future
network stress.

46
 Monitoring network and storage performance and utilization

Using performance data to detect congestion

The network fabric consists of the collection of components and their interconnections and interfaces that
make up the network. For the network to function at its theoretical maximum, all components must function
effectively, individually and together as a cooperative unit. Because Cisco UCS Performance Manager provides
a view into the network fabric, you can analyze performance data that network management tools cannot access.
This performance data encompasses components such as server ports, Ethernet links, FC uplinks, and so on.

Analyze data returned from these components to determine whether congestion exists, and what additional
effects the congestion is generating. Detect congestion by observing the performance data of specific network
components that show low throughput or exceptionally high usage rates. Low-performing or over-subscribed
components can contribute to or be the result of congestion.

Constraints are any limitations within the fabric that affect performance of the network, measured as throughput
or relative available remaining capacity. Constraints that affect data flow within the network can result
in network congestion and potentially network paralysis or collapse. These constraints can take the form
of component over-subscription, component misuse, inadequate provisioning or configuration, or failing
components or architecture.

Using network topology to review performance capacity

A quick method to determine how close your network is to maximum performance capacity is to display the
Topology view and analyze it for information about the connection quality between various components.

To display the Topology view, navigate to the Dashboard and click Topology. You can then click a
component, connection line, or an event icon for more information.

The Topology view includes graphical bandwidth usage information in the form of speedometer dials for both
overall and port bandwidth usage. It also provides a click-able diagrammatic map of device connections.

Figure 45: Topology view: bandwidth dials

Reviewing historical utilization and congestion

Cisco UCS Performance Manager uses historical log data to generate reports. Options for each type of report
refine the data.

To view historical information about interface utilization:

1 Navigate to Reports > Cisco UCS Capacity Reports > Interface Utilization.
In the right pane, the Report Parameters page appears.
2 In Root Organizer, choose /Devices/Network.
3 Select start and end dates.
4 Click Update.

47
Cisco UCS Performance Manager User Guide

In the lower section of the right pane, the Interface Utilization report displays historical data for the interface.

Figure 46: Interface Utilization report

5 For each interface, review the in and out average, and maximum and minimum values.
6 To identify interfaces that are causing or might cause network congestion, review the Forecasted In
Exhaustion and Forecasted Out Exhaustion values for each interface.
7 To display the page for a device or interface, click its name.

Viewing northbound port utilization by domain

The Network Capacity portlet provides bandwidth utilization for a specific domain from the fabric
interconnects to northbound LAN and SAN clouds.

The portlet displays ports that are operational and currently mapped to another device for a specific time range.
It shows the port name and role, the average Tx and Rx utilization, and event summary. You can add columns
for maximum Tx and Rx utilization.

1 Add the Network Capacity portlet to the Dashboard, as described in Adding a portlet to the Dashboard on
page 8.
The Network Capacity portlet appears on the Dashboard.

Figure 47: Network Capacity portlet

48
 Monitoring network and storage performance and utilization

2 To view information for a specific domain, in the upper right corner, click Domain and choose from the
drop-down list.
3 To change the time range, click Time Range and choose the range.
4 To add maximum utilization information, near the right side of any column, click the drop-down menu, click
Columns, and then click Max Tx and Max Rx.

Figure 48: Add Max Tx and Max Rx to Network Capacity portlet

5 For additional information about a particular port, click the port name.
The corresponding information page appears. The following example shows the Fibre Channel Ports page.

Figure 49: Fibre Channel Ports page

6 Review utilization graphs or use the Display drop-down menu to continue working with the selected
component.

Reviewing storage utilization and capacity

Cisco UCS Performance Manager monitors EMC storage devices, NetApp filers, and local storage (datastores)
of Hypervisor devices. To review storage capacity and utilization, use the following methods:

■ Review information for the device and its components by starting at the Infrastructure page.
■ Get at-a-glance information on the Storage Capacity portlet, and then focus on a device.

49
Cisco UCS Performance Manager User Guide

■ Generate a Storage Utilization vs Capacity report.

The procedures that follow describe how to use each of these methods.

Reviewing storage devices and components

Review storage capacity and utilization by drilling into the storage infrastructure. Based on your findings, you
can determine whether a resource is oversubscribed or experiencing other issues. You can then make necessary
adjustments to correct the problem. This procedure provides examples for an EMC device and a NetApp filer.

1 Navigate to the Infrastructure page.

2 In the left pane, expand the Storage organizer.
3 In the right pane, click a storage device.
The device's overview page appears.
4 In the left pane, drill down into the device's component list. For example:

■ For an EMC device, click Storage Pools.

The Storage Pools page includes a list of arrays with the corresponding graphs for the selected array.
The following example of the Space Utilization graph shows the current subscribed, managed, and
raw percentages for the selected storage pool. Raw represents the raw disk capacity minus RAID and
other overhead. In this case, the Subscribed value appears to be oversubscribed when compared to the
Managed and Raw percentages. You can address the Subscription percentage by adding drives to the
pool.

Figure 50: Space Utilization graph for EMC storage pools

■ For a NetApp filer, click Volumes.

The Volumes page includes a list of volumes with corresponding graphs for the selected volume. The
following example shows that vol3 is at 93% utilization, leaving 78.82GB available out of the total
1.10TB of space. The graph also depicts 93% current usage.

50
 Monitoring network and storage performance and utilization

Figure 51: NetApp Usage for a single volume

5 To review information and graphs for another component, in the left pane, select the component.
6 To see view multiple components of the same type on a single graph:
a In the left pane, click Component Graphs.
b At the top of the graph page, from Component Type, select an option.
c From Graph, select an option, and then click All on Same Graph.
The following example graph shows usage for all volumes on the same graph.

51
Cisco UCS Performance Manager User Guide

Figure 52: Usage graph for all NetApp volumes

d To hide or show data for a component, in the graph's upper legend, click the component name.
All on Same Graph provides a quick visual perspective on where potential storage issues exist and
where you have extra storage space that can help alleviate a problem.
e To display a list of dependents and dependencies for the selected storage component, click Dependency
View.

Reviewing hypervisor storage

Cisco UCS Performance Manager supports vSphere and Hyper-V devices, which provide local storage. Review
storage information to identify oversubscribed resources and those that have extra capacity.

1 Navigate to Infrastructure > Devices, and then select a virtual device:

■ For a vSphere device, select vSphere.

■ For a Hyper-V device, select Server > Microsoft > HyperV.
For the top ten datastores and hosts, storage and compute information appears.

52
 Monitoring network and storage performance and utilization

Figure 53: Storage graph for hypervisor

Viewing capacity utilization of storage devices

Before you can display data for storage devices (for example, EMC and NetApp) and the vCenter server
appliance, you must model the devices and monitor them. After you model and monitor these devices, storage
capacity information appears.
Use the Storage Capacity portlet to view capacity utilization of currently modeled and monitored storage
devices.

Figure 54: Storage Capacity portlet

The Storage Capacity portlet uses visual indicators as follows:

■ Color shows usage on an exponential curve, between green and red. Green indicates a usage of 50% or less,
going up the color spectrum to red, which typically indicates a usage greater than 90%.
■ Gradient becomes darker as the percent usage increases.
■ As the percent usage increases, the colored gradient bar lengthens.
■ Hover over a colored bar to display the component's actual percentage used.

The Client column displays the device that uses the storage component. N/A in the column means that no
devices are using that storage component, or the client device itself is not being monitored.

1 Add the Storage Capacity portlet to the Dashboard, as described in Adding a portlet to the Dashboard on
page 8.

53
Cisco UCS Performance Manager User Guide

2 By default, components with the highest LUN/Export, volume, and pool utilization are listed first. To view
a different perspective on storage capacity, click a column heading and choose Sort Ascending or Sort
Descending.
3 With the portlet added to the Dashboard, hover over an item to view a tooltip with current usage and
capacity information. For example:

Figure 55: Storage Capacity tooltip

4 To investigate a specific resource, click the client name hyperlink.

The component's page appears.

Figure 56: Datastores page

5 To continue working with the selected component, review the available graphs or select an option from the
Display drop-down menu.

Analyzing storage utilization versus capacity

Cisco UCS Performance Manager automatically generates the Storage Utilization vs. Capacity report when
you select the report. You can customize the report by setting report parameter options. Use information in the
report to determine whether to make adjustments to your overall storage utilization plan.
The Storage Utilization vs. Capacity report is only available with a full Cisco UCS Performance Manager
license.

1 Navigate to the Reports page.

2 Expand Cisco UCS Capacity Reports and click Storage Utilization vs. Capacity.

54
 Monitoring network and storage performance and utilization

The report using the default settings appears. The report contains information for the last seven days and
includes all currently monitored storage classes and structures.

Figure 57: Default Storage Capacity vs Utilization report

3 Review the current Report Parameter settings.

4 To review graphs available in the report, scroll down.
5 To print the report, click Printable.
6 To modify the report:
a Select start and end dates.
b Select options from the Report Parameter drop-down menus.
c Click Update.

55
Cisco UCS Performance Manager User Guide

Working with capacity projections and

finding available capacity 4
This chapter describes how to work with projected capacity exhaustion dates, trend lines, and predictive
thresholds.

Cisco UCS Performance Manager helps you determine where available remaining capacity exists in your
system. You can then alter resource allocation to make use of the available remaining capacity and alleviate
over-subscription to other components.

Usage and available remaining capacity data is available for individual ports, or collectively by role in the
Aggregation Pools view. Aggregation pools are logical bundles of multiple physical network interfaces within a
UCS domain, and are similar to network device port channels.

Viewing projected bandwidth exhaustion dates

From the Bandwidth Usage view, you can see the projected capacity exhaustion dates for both Rx and Tx. The
projected exhaustion dates alert you to a potential bandwidth issue before a threshold is crossed.

1 Navigate to the Infrastructure page.

2 Expand the CiscoUCS organizer and select a device.
The device's overview page appears.
3 In the left pane, click Components > Fabric Interconnects.
The Fabric Interconnects page appears.
4 In the left pane, click Bandwidth Usage.
The Bandwidth Usage page appears, and by default, lists the network-related components that belong to the
device.
5 To view the device's network-related components, beside Group By, click Network.
6 Expand the component names until you reach a fabric interconnect uplink.
7 To view the Projected Values in the Bandwidth Capacity Utilization graph, highlight the row of the
uplink.
If there is not enough collected data to make a projection, the projected values are zero.
8 To see Projected Tx Date and Projected Rx Date, use the horizontal scroll bar in the table.
9 Review graphs on the right side of the page. (The graphs vary by component.)
a To view the currently defined thresholds in a Bandwidth Capacity Utilization graph, click Projected
High Pct Tx* and Project High Pct Rx*.
b To move the graph along the time-line, click the right arrow.
c To view details for a particular data point, hover over the graph.

56
 Working with capacity projections and finding available capacity

Figure 58: Graph with projected exhaustion dates

d To export the graph data, click the gear icon and select Export to CSV.

Finding remaining capacity for components

You can determine available remaining capacity values for most system components. The following example
shows how to determine the remaining capacity for an Ethernet port.

1 Navigate to the Infrastructure page and expand the CiscoUCS container.

2 Select a device.
The device's overview page appears.
3 In the left pane, click Components > Ethernet Ports.
The page displays Ethernet ports information at the top and performance graphs at the bottom.
4 Click a port name that has a value listed in the Port Channel column.
5 Review graphs for bandwidth capacity and Ethernet port utilization.

57
Cisco UCS Performance Manager User Guide

Figure 59: Ethernet Port Utilization graph

Finding remaining capacity for an aggregation pool

An aggregation pool is a logical bundling of multiple physical network interfaces, commonly known as a
port channel. For example, the Per Chassis Ethernet Pool includes all links from all chassis to all fabric
interconnects, and is used for chassis bandwidth balance comparison. The following example shows how to find
remaining capacity for an aggregation pool.

1 Navigate to the Infrastructure page and expand the CiscoUCS organizer.

2 Select a device.
The device's overview page appears.
3 Click Components > Aggregation Pools.
The page displays aggregation pool information at the top and performance graphs at the bottom. .
4 Review the Per Chassis Server Port Bandwidth Capacity Utilization graph.

58
 Working with capacity projections and finding available capacity

Figure 60: Per Chassis Server Port Bandwidth Capacity Utilization graph

For more information about aggregation pools, refer to the Cisco UCS Performance Manager Administration
Guide.

Creating a predictive threshold

A threshold defines a value that a data point should not go beyond. When a threshold is reached, Cisco UCS
Performance Manager generates an event. You can customize an existing default threshold or create a new
threshold.
The following procedure shows you how to create a predictive threshold for a chassis.

1 Modify the graph template:

a Navigate to the Infrastructure page and expand the CiscoUCS organizer.
b Select a device.
The device's overview page appears.
c Click Components > Chassis, and then select a chassis.
d From the Display drop-down menu, choose Templates.

Figure 61: Display menu

e From the Template drop-down menu, choose UCSCapChassis.

59
Cisco UCS Performance Manager User Guide

Figure 62: Template menu

2 Create a new threshold:

a To the right of Thresholds, click +.
b Enter a name for the threshold; for example, Throughput Bits Rx.
c For the type, select Predictive Threshold.

Figure 63: Add Threshold dialog box

d Click Add.
The new threshold is added to the Thresholds table.
3 Edit the threshold values:
a Double-click the new threshold name.
The Edit Threshold dialog box appears.
b From the DataPoint drop-down menu, select throughputRx_bitsRx.
c In the Minimum Value and Maximum Value options, enter numeric value.

Figure 64: Edit Threshold dialog box

d Click Save.
The new predictive threshold is added to the device.

60
 Working with capacity projections and finding available capacity

Adding a trend line to a graph

Trend lines show the projected utilization of a device or component over a specified time range.
Before completing this procedure, create a new predictive threshold, as described in Creating a predictive
threshold on page 59.

1 Navigate to the Infrastructure page and expand the CiscoUCS organizer.

2 Select a device.
The device's overview page appears.
3 Select Components > Chassis, and then select a chassis.
4 From the Display drop-down menu, choose Templates.
5 Create a new graph, or choose an existing graph.
6 From the action drop-down menu, select Manage Graph Points.
The Manage Graph Points dialog box appears.
7 Click + > Threshold.

Figure 65: Manage Graph Points

8 Select the threshold (for example, Throughput Bits Rx), and then clickSubmit.
9 In the Manage Graph Points dialog box, click Save.
10 To view the trend line, change the Display drop-down menu to Graphs, and scroll down to the graph.

Note By default, the trend line calculation includes only the last ten days of data. There might not be
enough data to see the trend line appear on the graph. However, you can view the projected exhaustion dates
in the lower right corner.

61