Improve the methods which delete older log files when log partition usage meets the defined thresholds
Currently 3 thresholds can be configured to take various actions dependent on log disk space usage. When set to delete older log files, let's say - at 85% - this simply executes the logcleaner.plx job at 01:30am each morning. If the threshold is say 80% at this time, and subsequently lots of activity of the box, (typically abnormal activity) and the log partition fills up before the end of the day to 100%, then the raw log files fail to be compressed down by archiving jobs and stay at their current size - logging stops as syslog has nowhere to write, Postgres cannot write also, and when the logcleaner job runs on subsequent days it doesn't effectively delete the old log files to fix itself. I'm aware this is because the system is by then in an error state, but - it would be good if we could look at a more regular method of checking the current log disk size and clearing it if necessary to stop these issues happening. Even something as simple as having the job run more regularly (every 4 or 6 hours?) would typically prevent this from happening. To fix the issue when in this error state you must manually make space on the partition in question.