Hi all
From time to time, I see servers struggling with low memory or high cpu usage without zabbix showing the cause. As with memory, it might be a php memory leak or something and as for cpu, perhaps some other bug.
Mostly, these issues are memory leaks, but then, I don't know where to start. It may be apache systems or nginx og whatever webserver was chosen.
Could it be possible to use the AI things in current Zabbix to track down processes that deviate from normthal? I mean - sshd usually just stays the same and just forks out when someone logs in and the number of logged in users is easily monitored. What's hard, is to find, in general, what is using the system's memory? By tracking each PID, it would perhaps be possible, but then, after a reboot, that PID won't' match the old one etc.
So - any idea how to make a generic test for this? Atm I have at least two servers misbehaving like this - both linux.
Thanks
roy
From time to time, I see servers struggling with low memory or high cpu usage without zabbix showing the cause. As with memory, it might be a php memory leak or something and as for cpu, perhaps some other bug.
Mostly, these issues are memory leaks, but then, I don't know where to start. It may be apache systems or nginx og whatever webserver was chosen.
Could it be possible to use the AI things in current Zabbix to track down processes that deviate from normthal? I mean - sshd usually just stays the same and just forks out when someone logs in and the number of logged in users is easily monitored. What's hard, is to find, in general, what is using the system's memory? By tracking each PID, it would perhaps be possible, but then, after a reboot, that PID won't' match the old one etc.
So - any idea how to make a generic test for this? Atm I have at least two servers misbehaving like this - both linux.
Thanks
roy