The Experience Analytics reporting system is driven by data collected from the Experience Analytics Processing Servers. Collected data consists of statistical information that is generated while sessions are processed. The data is generated in one-minute buckets and extracted into the reporting database every five minutes.
The actual text of the sessions remains on the Processing Server and is not migrated to the reporting databases.
The collected data is aggregated into hourly and daily reporting data.
After data has aged a pre-defined period of time, the data is removed from the database so that it can be kept to a manageable size.
The length of the retention period is positively correlated to the size of the database; retaining more data can result in a very large database, particularly in the tables that store hourly data.
Session data retention
Session data is stored on the local disk for each Processing Server (Canister) in your environment. It is not retained in a Experience Analytics database. Configuration of session data retention is managed for each Canister through TMS.
Data collection is enabled and disabled through the
Data Collection setting.
Note: This setting should not be disabled during normal operations. It is available primarily for debugging purposes.
Data Collector Logging Level: Specify the logging level for the Data Collector only. Don't change this setting unless directed by Experience Analytics. Levels:
- 0 -
- 1 -
- 2 -
- 3 -
- 4 -
- 5 -
- 6 -
- 7 -
Note: Status level messages always appear in the log for any non-zero logging level.
This value overrides the system logging level, which can be configured through TMS.
Managing data aggregation performance
Depending on the volume of data retained in your reporting database, the periodic data aggregation runs performed by the Data Collector can impact performance and consume unnecessary space. Performance in the Report Builder can also be impacted.
To improve performance, Experience Analytics provides the following controls to define the frequency and time of day that the hourly data is aggregated to the daily level.
Data Aggregation - Daily Data Processing
- Describes the available settings for the date range over which daily data aggregations occur. Options:
Hourly through Current Hour- Aggregate once per hour, in first data collector run of the hour, using most current available hourly data
Daily Through Previous Day- Aggregate daily data once per day, using hourly data up through the previous day
Daily Through Start of Hourly Retention Period- Aggregate daily data once per day using hourly data up to the beginning of the hourly data retention period
Note: The default value,
Daily Through Start of Hourly Retention Period, optimizes the aggregation process for performance of data collection. Configuring this setting to a value other than
Hourly Retention Periodretains overlapping aggregated data in the database and may impact system performance during data aggregation.
Data Aggregation - Daily Data Time of Day
- This setting defines the hour of the day when the daily data aggregation run is performed, if the daily data aggregation is set to occur on a daily basis. By default, it is configured to run at
2:00in the Experience Analytics system timezone.
Note: Configure the daily data aggregation run to be performed during an off-peak hour.
Another method of improving the data aggregation performance is to install a data preaggregator. A data preaggregator aggregates data before it is sent to the data collector, which reduces the amount of processing that occurs on a data collector.
The Experience Analytics reporting system is driven by event-based data. All reporting data is based on the defined events and dimensions and is stored in the reporting database:
The data is divided into two types.
Non-dimensional event data. This data consists of statistics and values gathered for events without respect to any dimension data.
Dimension-based event data. This data is based on the defined report groups assigned to an event. Each report group may contain multiple dimensions along with the timestamp and corresponding event counts and values.