Summary
This document describes how to effectively manage the Citrix EdgeSight database grooming schedule.
Background
Citrix EdgeSight collects a wide range of performance, availability, and usage data about end-user systems, applications, and the network. The Citrix EdgeSight Agents collect data from end-user systems and upload it to a Citrix EdgeSight Server. Depending on the number of end-user systems, applications, and network activity, databases can grow quickly without proper management. The primary database management mechanism is grooming.
Grooming is the process of removing older data from a database at regular intervals to make room for new data. Grooming is critical for maintaining efficient database operation. An effective grooming schedule controls database size and helps ensure acceptable performance while retaining sufficient data for business operations.
The following example shows how database grooming settings affect the size of the Citrix EdgeSight Server database. You have deployed Citrix EdgeSight agents on 2500 end user devices. The devices are running an average of 50 processes and visit an average of 100 Web sites over the period of time that network data is retained. The agents are collecting data for 12 hours each workday. Changing the grooming parameter for Network Statistics from 7 to 14 days increases the size of the database by about 40 percent, roughly equivalent in its effect on database size to adding 1000 devices. This illustrates the importance of establishing an effective grooming policy.
Grooming Schedule
Citrix EdgeSight has a distributed structure, with Citrix EdgeSight Agent databases on each managed device which are uploaded to a single Citrix EdgeSight Server database. The data retention settings for the agent database are specified as part of the agent configuration and are applied to devices based on their department membership.
The grooming schedule for the server database is specified as part of the server configuration. The server grooming schedule allows you to specify the number of days that data is retained by data type. This allows data used to identify trends, such as performance data, to be retained longer than data which quickly becomes stale, such as real-time alert data.
The default values for grooming the server database are sufficient for most installations. You may want to use the default values at first and adjust them over time based on user requirements and system performance. In cases where you want to retain more data, consider creating an archive report or performing data warehousing as methods of keeping historical data in preference to relaxing the grooming configuration.
Server Database Grooming
To edit the server grooming schedule, open the Citrix EdgeSight Server Console and navigate to Server Settings > Maintenance > Grooming > Configuration. The Grooming table contains the following information:
• Data—The type of data to be groomed.
• Table—The database table where the data is stored.
• Default Days—The default number of days that data of the selected type is retained.
• Current Days—The current number of days after which data of the selected type will be groomed.
The following table shows the default grooming schedule for Citrix EdgeSight Server data.
Data | Table | Default Days |
Application Performance | image_perf | 30 |
Application Usage | usage | 90 |
Citrix Session Performance Data | ctrx_session_perf | 30 |
Citrix System Performance Data | ctrx_system_perf | 30 |
Company Instance Records | instance | 30 |
Disk Usage | drive_space | 30 |
Historical Alerts | alert | 30 |
Incoming Alerts | alert_incoming | 3 |
Light Trace Performance Data | ltrace_event | 30 |
Messages | message | 30 |
Network Statistics | core_net_stat | 10 |
Network Transactions | core_net_trans | 10 |
System Performance Data | system_perf | 30 |
In most cases, the grooming schedule is configured to retain one month of data. Application usage data is retained for 90 days due to the need in many environments to track application usage for license and compliance reporting. On the other hand, network data is only retained for 10 days due to high data volumes and the transient nature of the data.
The grooming strategy for a specific data type should take into account how fast the usefulness of the data decreases from time of collection and also how much data is collected on average over a time period. For example, the data in the alert_incoming table has a short shelf life. Real-time alerts are intended to address critical problems that can be resolved by taking action within a short timeframe, such as crashes of mission-critical applications or disruptive network failures. Because of these characteristics and because historical alert data is retained, real-time alert data is groomed aggressively.
It is important to ensure that the grooming schedule is taken into account if data is being warehoused or reports are being archived. If data is transferred less frequently than the grooming schedule for a type of data, data loss can occur. Similarly, report archiving schedules must take into account the grooming schedule to avoid introducing gaps in historical reports.
You can monitor the status of grooming jobs by displaying the grooming log. The log displays the following information:
• Data Area—The type of data on which grooming was performed.
• Grooming Job Name—The name of the grooming job run.
• Job Status—The completion status of the job.
• Start Time—The date and time that the grooming job started.
• Duration—The elapsed time taken by the grooming job. Note that in the case of smaller databases, grooming jobs may show a duration of zero time.
Agent Data Retention Settings
You may also need to manage agent database size. The agent data retention parameters are defined in the agent configuration. Data is retained in the agent database according to the following configuration values:
• Days To Keep In DB—The number of days that data is retained in an agent database that has been uploaded to the server. After this number of days has elapsed, data that has been uploaded is groomed from the agent database. The default value of 8 days guarantees a week of data available on a device which uploads regularly.
• Max Days To Keep In DB—The number of days that data is retained in an agent database that has not been uploaded to the server. After this number of days has elapsed, data is groomed from the agent database whether or not it has been uploaded to the server. The default value of 29 days helps ensure that devices which have been disconnected from the network have the opportunity to upload data.
Best practice dictates that these values not be increased; increasing these parameters can have a negative impact on end-user system performance. In some cases, these values should be decreased in order to accommodate slower machines or machines accommodating a large number of concurrent users and processes.
Also, increasing data collection settings in the agent configuration may require stricter data retention settings. Configuring longer collection times can result in more data in the database which, in turn, may require more aggressive grooming to maintain optimal database size and avoid system performance issues.
