Comment 3 for bug 1584093

I can confirm the problem, and it seems it is due to us reading syslog and syslog
getting constant updates. ( tail -f /var/log/syslog).

The script needs to parse every syslog line for potential apparmor errors which is
expensive. There was another issue with listening on the pipes but it alone did not
solve the problem. Only completely disabling the parsing of syslog updates stops python
from taking up CPU.

The question is now do we want to drop reading syslog files or do we live with the CPU
consumption...

Another question is though why syslog is flooded with messages.