Citrix Director: Trends Explained

Trends page in Citrix Director gives an overview of the different parameters of the Citrix XenDesktop/XenApp deployment over time. This helps in analyzing and predicting issues. Below, are the different reports in the Citrix Director Trends view.

The Sessions tab provides a report of the peak number of sessions across delivery groups for a selected time period.

You can toggle between different Delivery Groups as well as filter the time period for which you want the Trends report. The time period provides a selection of the last 2 hours, last 24 hours, last week, last month, last year, and a custom time period.

The graph shown on the panel after you select the above two criteria, shows three important parameters.

1. Peak Connected Sessions – The maximum number of connected sessions, both desktops and apps at a given point of time.
2. Peak Disconnected Sessions- The maximum number of sessions, both desktops and apps that were earlier active but are now at a disconnected state, at that point of time.
3. Peak Concurrent Sessions – The maximum number of registered sessions, connected and disconnected.

The session details table provides an in depth view about the session details that were accounted for during the time period and filter selected, its users and the associated delivery groups.

The session details table shows the following parameters.

1. Associated Users –  The users associated with the sessions. The associated users can also be searched and filtered by using the filter box above.
2. Machine Name – The name of the machine running the session.
3. Delivery Group – The name of the delivery group hosting the session.
4. Session Start Time – Time when the session started, without considering the time for reconnection if any
5. Session End Time – Time when the session ended.  Note – the session may still be running at which time “Running” will be displayed in this field.
6. Session Duration – Time difference between the start of the session till it ends in hours and minutes.

The connection failure tab gives an overview of the different types of connection failures that have occurred across different Delivery Groups. This helps in capacity management and addressing contingency strategy.

We can search and narrow down the connection failures by choosing the machine type, failure type and then the Delivery Group and time period.

The machines can be filtered by one of the following types:

1. Desktop OS
2. Server OS

The Failure Types can be filtered the below types:

1. Client Connection Failures – Failures due to the inability of the client side to complete the session connection. For example, connection timed out, server was not reachable
2. Configuration Errors – Failures caused due to configuration errors.  For example, an administrator put a Delivery Group or a particular machine in Maintenance mode
3. Machine Failures – Connection failures as a result of the machine in failed state or failure to start up or respond
4. Unavailable Capacity – Failures due to the configured capacity of a particular delivery group having been completely consumed. For example, too many users logged into a Server Desktop OS delivery group or a user accessing a Pooled Random delivery group once all the machines in the delivery group are already assigned to other users
5. Unavailable Licenses – Failures when the delivery controller is unable to acquire a license from the license server to launch the session

Failure reasons for the Failure Types are shown in the Session Details table:

1. Session Preparation – Session prepare request from Broker service to VDA failed
2. Registration Timeout – VDA did not register in time after being powered on for session launch
3. Connection Timeout – Client did not connect to VDA after VDA was prepared for session launch.
4. Licensing – License request failed (typically no license available).
5. Ticketing – Failure during ticketing, indicating that the client connection to VDA does not match the brokered request
6. Other – General unresolved errors between client and VDI
7. General Fail – General unresolved errors during initial brokering operation
8. Maintenance Mode – Machine or delivery group in maintenance mode
9. Application Disabled – Application can no longer be used and has been disabled by the admin
10. License Feature Refused – Feature being used is not licensed (e.g. license doesn’t allow use of multi-session VDAs)
11. Session Limit Reached – No more sessions allowed (e.g. launching second session on VDI VDA)
12. Disallowed Protocol – Resource doesn’t allow protocol requested by session launch
13. Resource Unavailable – Resource has been removed/disabled since enumeration
14. Active Session Reconnect Disabled – Session is already connected to a different endpoint but session stealing is disabled
15. No Session To Reconnect – Specified session could not be found during reconnect
16. Spin up Failed – VDA could not be powered-on for session launch
17. Refused – VDA actively refused session prepare request
18. Configuration Set Failure – Failed to send required configuration data to VDA prior to session launch
19. No Machine Available – No machine available for launch (e.g. all shared VDI machines in group are in use).
20. Machine Not functional – Non-power-managed assigned machine not registered

Similar to the Sessions panel, the Connection Failure panel shows trending of connection failures that have occurred in the selected site. The Failed Connections Details table includes associated users, machine name, and delivery group. Apart from these it also displays:

1. Failure Time – The time stamp when the session failed.
2. Failure Type – The type of failure which is explained in detail above.
3. Failure Reason – Connection Failures

The failed desktop OS machines tab gives an overview of the different problems associated with failures in desktop machines.

The trends for failed desktop OS machines can be filtered by the different failure types which includes:

1. Failed to Start – Machine failed to start up when initialized.
2. Stuck on Boot – Broker asked VM to boot, and <n> seconds have elapsed since the Broker had asked the VM to boot. VM has not registered since the time it was asked to boot.
3. Unregistered – When the desktop is still unregistered with the controller.

The graph gives the failures and events associated with the filter criteria. The machine failure details table gives information about the associated machine, the delivery group, the failure time, type and reason. Machines names can also be searched using the search box provided.

The failed server OS machines tab gives information on the servers that have failed to provision and those that are on heavily loaded.

Failed server OS machines can be classified on their failure types:

1. Failed to Start – Provisioned server did not start up.
2. Stuck on Boot – Provisioned server failed to boot successfully..
3. Unregistered – VDA in an unregistered stated
4. Maximum Load – Machines that are at maximum defined load via the defined load evaluators

The graph gives a clear picture of the failure events over time.

The server machine failure details page gives a detailed picture and is similar to the desktop OS failure details.

A very important feature which helps us to analyze the logon performance of the users who have connected to the XenDesktop/XenApp deployment through trending and baseline analysis.  This enables customers to compare the average logon duration, broken down into steps and compare to previous timeframes based on the customers’ previous logon performance in that environment.

The filter criteria includes Delivery Group and time period to zero in on the required trend.

The graph includes the following important parameters in analyzing the performance of the VDI deployment.

1. Brokering: The time taken to complete the process of brokering the session.
2. VM Start: In case the session required a machine to be started, the time taken to start the VM.
3. HDX Connection: The time taken to complete the steps required in setting up the HDX connection from the client to the VM.
4. Authentication: The time taken to complete authentication to the remote session.
5. GPOs: In case any Group Policy settings have been enabled on the machines, the time taken for the GPOs to be applied.
6. Logon Scripts: In case logon scripts are configured for the session, the time taken for the logon scripts to be executed.
7. Profile Load: In case profile settings are configured for the user or the machine, the time taken for the profile to be loaded.
8. Interactive Session: The time taken to handoff keyboard and mouse control to the user.

Note: Total Logon Duration steps run asynchronously with only the primary steps being shown.

Load evaluator index is a feature that was used in XenApp.  The evaluator has changed in XA 7.x to more accurately reflect the important metrics.  This feature helps in understanding the resource utilized by the XenApp deployment over time.

The graph shows trends which include CPU, memory, disk and session count details.

With XenDesktop 7, Citrix introduces Load Management which is different in functionality that administrators are familiar with in XenApp. A server’s load index may be the aggregate of:

• Various computer performance counter based metrics, namely CPU, Memory and Disk Usage

• Session Count

It is designed to indicate how suitable a XenApp Worker is to receive a new user session. It is the Delivery Controller’s responsibility to calculate the load index based on the aggregate of the normalized load rule indexes generated by the various load rules. As only the Delivery Controller can determine the session load, a server’s overall load index is calculated on the Delivery Controller and not the Virtual Delivery Agent.

In Director, the Load Evaluator Index is shown as compared to the number of connected users on the XenApp delivered application server. Administrators are able to choose which metric (Total, CPU, Memory, Disk, Sessions) to compare to the connected users. Each data point in the trend has the total number of users connected to RDS Workers for the last interval. In Director, the Load Evaluator Index is converted to a percentage (loadeval/100) and reflected in the graph (for charting purposes only).

The hosted application usage analytics provided by the Citrix Director has usage details about:

1. All hosted applications – Details about all the applications published in the site
2. Single applications – Drill down information about each of the applications in detail

Note: This feature is available in the platinum license edition

The graph gives a picture of the ‘Peak Concurrent Application Instances’, which is the maximum number of application instances at that given point of time.  Each data point shows the max number of application instances running concurrently for that data point interval.

Note: 2 hours filtering criteria is not available

1. Application Name – Name of the application
2. Folder Path – Path where the application has been executed from
3. Peak Concurrent Instances – Maximum number of application instantiated at that point of time
4. Total Launches – Number of times the app has been launched
5. Total Usage Duration – Number of hours and minutes the application has been in use
6. Total Distinct Users – Number of distinct (unique) users who have used the application

1. User Name – Name of the user
2. Total Instances – Total instances of the application at that point of time
3. Total Launches – Total number of times the application has been launched
4. Total Usage Duration – Total usage of the application by the user in hours and minutes
5. Average Usage Duration – Average usage of the application, calculated from the ‘Total Usage Duration’ and  ‘Total Launches’

Note: The  columns are sortable.  The table will list applications that have not been used during that time period for the purpose of showing the most/least used applications. The admin is able to click on a specific application from the table and drill down to user specific details for that application.

The network analytics provided by the Citrix Director is through the Netscaler HDX Insight.

Note: More information about Network Insight can be found at the below blog post.

/blogs/2013/03/19/netscaler-insight-center/

/blogs/2013/09/19/hdx-insight-integration-with-director/