You can configure the behavior of each QueueSink with the schedule and dataItemThreshold properties. The schedule property sets a time schedule for flushing the queue and writing all the events in the queue to a log. The dataItemThreshold sets a maximum number of events to accumulate in the queue before writing all the events in the queue to a log. The queue is flushed when it either meets the schedule or reaches the item threshold. Both of these properties are set centrally for all QueueSinks in the SQLFlushSchedule and dataItemThreshold properties of the /atg/reporting/Configuration component. You can also customize each QueueSink with its own properties.
FileSinks
The FileSink components are atg.service.datacollection.FormattingFileLogger class components that log events to a flat file. A FileSink has properties that specify the directory and file name prefix of the event log. For example, the requests/FileSink component has these properties:
logFileDir^=../Configuration/dataLogFileDir
logFileName=request_
Other properties of the FileSink components refer to the /atg/reporting/Configuration component to specify the exact format of the log file and log file name and the log’s rotation schedule. By default, the log files include timestamps in their file names, so you may see a user events log file name like this:
userevents_02-09-2001_18-36-03-55.data
For more information about logging services, see the Logging and Data Collection chapter in the Platform Programming Guide.
SQLSinks
The SQLSink components are atg.service.datacollection.SQLTableLogger class components that log events to a SQL database table. A SQLSink has two properties that specify the database table and column mapping to use when logging events to the database. For example, the requests/SQLSink component has these properties:
tableName=dps_requests
SQLColumnMappings=id:id,timestampAsDate:timestamp,\
sessionId:sessionid,name:name,memberAsDBValue:member
The SQLSink also has a read-only property that sets the SQL INSERT statement to use when writing log events to the database.
