I'm personally beyond logrotate and have the opinion that whatever we do on this level is just a bandaid. I'm looking for a way not even to accumulate logs in container root disks and do stuff like writing the DHCP agents files out to a bind mount, that alone would have fixed this particular issue, just from the customer impact. I'm not a friend of running logrotate every 5 minutes or so to reduce the likelihood that conntrack fills up the root disk because of a DDOS or whatever.
I'm personally beyond logrotate and have the opinion that whatever we do on this level is just a bandaid. I'm looking for a way not even to accumulate logs in container root disks and do stuff like writing the DHCP agents files out to a bind mount, that alone would have fixed this particular issue, just from the customer impact. I'm not a friend of running logrotate every 5 minutes or so to reduce the likelihood that conntrack fills up the root disk because of a DDOS or whatever.