This happens to be one of the areas where the production/stage setup is different from our local setups. On prod/stage, PHP sessions are stored in memcached. What that does is that it makes it impossible to store medium - large chunks of data in the session. Doing that is anyhow not advisable.
When an attempt is made to write a large piece of data to the session, it results in a write failure from php to memcached. Subsequently, a read failure from memcached immediately follows.
In cases where session information is needed, this results in irregular behavior. In our case, it resulted in content console users being logged out. The failure can be different in different cases but the pattern will be the same.
Check: PHP error logs, SELinux context (if SELinux is enabled) etc.
Clues: Failures occur more irregularly when there is a write failure as discussed. The failures are more regular with SELinux issues.
The fix for the SELinux issue is simple. Just check:
# getsebool httpd_can_network_memcache
That should output:
httpd_can_network_memcache --> on
If it outputs an off, just do:
# setsebool -P httpd_can_network_memcache 1
As simple as that.