When using the ATP: Endpoint feature and updating to ATP version 2.2, the web interface becomes unavailable 1-2 weeks after the update. If you access the command-line interface and run the 'df' command, you also see that /var is 100% utilized.
Here is an example of the output of df -h when this condition is occuring:
Filesystem Size Used Avail Use% Mounted on
/dev/sda7 20G 2.9G 16G 16% /
devtmpfs 16G 0 16G 0% /dev
tmpfs 16G 152K 16G 1% /dev/shm
tmpfs 16G 18M 16G 1% /run
tmpfs 16G 0 16G 0% /sys/fs/cgroup
/dev/sda1 477M 93M 355M 21% /boot
/dev/sda6 80G 572M 76G 1% /tmp
/dev/sda8 77G 73G 0 100% /var
none 2.0G 0 2.0G 0% /var/symantec/sgs-td/pie/tmp
/dev/sda3 110G 1.8G 103G 2% /var/lib/elasticsearch
/dev/sda5 96G 487M 91G 1% /var/log
/dev/sda2 110G 61M 105G 1% /var/backup
tmpfs 3.2G 0 3.2G 0% /run/user/994
tmpfs 3.2G 0 3.2G 0% /run/user/1002
tmpfs 3.2G 0 3.2G 0% /run/user/500
tmpfs 3.2G 0 3.2G 0% /run/user/1000
tmpfs 3.2G 0 3.2G 0% /run/user/1001
Some Endpoint customers were experiencing an excessive production of reputation request events which flooded the system. The web interface becomes inaccesible when these queries are queued into a file that fills the partition where the database that handles UI functions in stored in.
This issue has been resolved in ATP version 2.3. We resolved the source of the excessive event generation as well as improved the system’s resilience in environments with unexpected load.