Page Comparison

The new FileWriter uses the IValueSerializer API to serialize files. It is available since Kieker 1.14. Like the old file writers, it comes with a wide range of settings. It usually writes two files. One containing a map of strings to ids and one logfile.

To use the new FileWriter set

kieker.monitoring.writer=kieker.monitoring.writer.FileWriter

...

kieker.monitoring.writer.filesystem.FileWriter.customStoragePath=

The FileWriter uses UTF-8 as default char set. However, you may want to specify another charset with

kieker.monitoring.writer.filesystem.FileWriter.charsetName=UTF-8

As host file systems have limits on file length and to avoid losing all data when the log is corrupted, you can limit the maximal number of entries (events) per created file. The value must be greater than zero.

kieker.monitoring.writer.filesystem.FileWriter.maxEntriesInFile=25000

The maximal file size of the generated monitoring log. Older files will be deleted if this file size is exceeded. Given in MiB. Note: At least one file will always remain, regardless of size! Use -1 to ignore this functionality.

kieker.monitoring.writer.filesystem.FileWriter.maxLogSize=-1

The maximal number of log files generated. Older files will be deleted if this number is exceeded. At least one file will always remain, regardless of size! Use -1 to ignore this functionality.

kieker.monitoring.writer.filesystem.FileWriter.maxLogFiles=-1

As mentioned before, the FileWriter writes a map of all strings to a map file. This is usually done by the TextMapFileHandler, which is the default. You may choose another map file handler.

kieker.monitoring.writer.filesystem.FileWriter.mapFileHandler=kieker.monitoring.writer.filesystem.TextMapFileHandler

When flushing is disabled, it could require a lot of events before finally any writing to the map file is done. In case of long running observations, this is the desired behavior. However, in shorter experiments and in cases when the application crashes, it is helpful to ensure all map entries have been written as soon as possible. To force flushing on the the map file, set the following property to true.

kieker.monitoring.writer.filesystem.TextMapFileHandler.flush=true

As map files can become quite large, you may want to compress the output of the TextMapFileHandler (compression not supported yet)

kieker.monitoring.writer.filesystem.TextMapFileHandler.compression=kieker.monitoring.writer.filesystem.compression.NoneCompressionFilter

Log file pool handler manages when files are written, how they are named and when they are removed. The default RotatingLogFilePoolHandler which supports a upper limit of log files which are kept.

kieker.monitoring.writer.filesystem.FileWriter.logFilePoolHandler=kieker.monitoring.writer.filesystem.RotatingLogFilePoolHandler

The log stream handler writes the text output. Default is the TextLogStreamHandler (text serialization in standard Kieker format)

kieker.monitoring.writer.filesystem.FileWriter.logStreamHandler=kieker.monitoring.writer.filesystem.TextLogStreamHandler

Alternatively, you may use the BinaryLogStreamHandler which serializes the data in binary format.

kieker.monitoring.writer.filesystem.FileWriter.logStreamHandler=kieker.monitoring.writer.filesystem.BinaryLogStreamHandler

You may want to write your own LogStreamHandler, e.g., to support JSON as output format.

When flushing is disabled, it could require a lot of events before finally any writing to the log file(s) is done. To force flushing on the set the following parameter.

kieker.monitoring.writer.filesystem.FileWriter.flush=false

When flushing is disabled, records are buffered in memory before written. This setting configures the size of the used buffer in

bytes.kieker.monitoring.writer.filesystem.FileWriter.bufferSize=8192

When compression is enabled, each log file is written as zipped binary file.

kieker.monitoring.writer.filesystem.FileWriter.compression=kieker.monitoring.writer.filesystem.compression.NoneCompressionFilter

Depending on the libraries used alongside Kieker, you can user ZIP, GZIP and XZ compression. The corresponding classes are:

• NoneCompressionFilter no compression
• GZipCompressionFilter
• XZCompressionFilter
• ZipCompressionFilter

You can write your own compression filter utilizing the ICompressionFilter interface.