Here is the information that I was able to fine pertaining to how the logs are managed:
"Because log file sizes and their number grow over time, the Agent purges logs files when:
To prevent unnecessary delays in your application's start up, the Agent's bulk purge limit is set to delete a maximum of 50 of the oldest files. This logic is shared between bootstrap and core Agents, so in total 100 of the oldest files may be deleted at one time."
For more information, please refer to this link:
Hope that helps,
Hi Oleksandr M.,
The default log level (if you simply commented out the line in the configuration file) should normally be 'info'. Info would normally write less data to disk than a level of fine.
none | off, finest, finer, fine, info, warning, severe."
Also to add to David's answer, you should also note the particular details for agent log rotation from that link shared by David:
"Log rotation and aging prevent the Agent from filling up the host's file system.
Each Agent limits its number of logs to 5, rotating the log file
index. Because Index 0 contains environment and start up information, it
is never overwritten. Each of these 5 files is limited to whatever
logfilesizeis configured as <option> passed to the JVM as described above."
If I am not mistaken, the default value for logfilesize is 10MB. So typically, for a single agent it would be 5x 10MB = 50MB for the agent log. You also need to consider the bootstrap log and so that comes to a total of 100MB (for 2x 50MB).
Typically, those other limits quoted by David Nicholls above would start to come in once you have multiple agents per host (more agents = more log files), or the log file has aged (eg: even if individual log file size < logfilesize; it may still be purged due to the 14 day retention period)