Log File Rotation Frequency
Log files can be set to rotate by size or by period. These settings are processed once a day. However they need to be checked more frequently due to exceptional events such as brute force attacks.
Assume your settings are rotate by size at 500K. When the maintenance script runs if you a slightly under it will leave it alone, so maybe the following day it might reach 900K. So far so good.
But what happens when the log file grows to 14GB in 30 minutes, exhausting all free disk space and crashing the server. No harm was done by the brute force attempts, the damage was done by logging them.
We need an additional supervisor script that is running either continuously or with a frequency of say 5 minutes to look for exceptional log file sizes (adjustable threshold but say 10x setting), so that if this filesize is exceeded then log files are instantly stopped/deleted so we don’t fill our disks up to the brim, and a warning is issued to the admin.
The script needs to be nimble, so I suggest that all it does it check file sizes (server-wide, or per domain) and doesn’t waste cpu/resources by executing log file rotation... it is there simply to safeguard disk space
This is a valid request, so we’ll look into it. There is no ETA at the moment, but we would really appreciate you voting for this request so that we can accurately assess its popularity relative to other features. Thanks in advance!
Costis Papadakis commented
Email logs must be preserved for more days
Deepak Sakhrani commented
Email Logs are extremely necessary to track history of events. We need more details of past dates for review of spam / email delivery. It is an important feature so I vote for this feature.
Miguel Monteiro commented
Store email logs (pop logs, imap logs, smtp logs, etc) more than 24 hours on Windows.