Tracking and reporting the number and total size of log entries that are created, and then flushed to the disk, will help diagnosing production incidents where too much logging is causing availability issues. We may use this information to change the verbosity of logging at runtime.