Case in short:
I'm getting randomly false positive problems from switchport utilization. Item value could be 10Tbps on 100M port. I'm using 1m polling cycle.
I have been looking the same oid with snmpget from switch. Counter values seems to grow steady, but still Zabbix gives enormous values sometimes.
Items has not anything special... If[in|out]hcoctects with 2 preprocessing steps: change per second and custom multiplier 8.
After the peak one polling cycle is always lost. Like this:
00:00.10 5000 <normal value>
00:01.10 5000000000
00:02.10 <lost>
00:03.10 4550 <normal value>
Setup:
Debian 10 VM, 2C/4GB RAM
Zabbix 5.0.10
MySQL
~100NVPS
Any thoughts what could cause this?
I have seen some slow queries from zabbix_server log. After optimizing of history tables, slow queries does not appear anymore. Could it still be performance issue with database?
I'm getting randomly false positive problems from switchport utilization. Item value could be 10Tbps on 100M port. I'm using 1m polling cycle.
I have been looking the same oid with snmpget from switch. Counter values seems to grow steady, but still Zabbix gives enormous values sometimes.
Items has not anything special... If[in|out]hcoctects with 2 preprocessing steps: change per second and custom multiplier 8.
After the peak one polling cycle is always lost. Like this:
00:00.10 5000 <normal value>
00:01.10 5000000000
00:02.10 <lost>
00:03.10 4550 <normal value>
Setup:
Debian 10 VM, 2C/4GB RAM
Zabbix 5.0.10
MySQL
~100NVPS
Any thoughts what could cause this?
I have seen some slow queries from zabbix_server log. After optimizing of history tables, slow queries does not appear anymore. Could it still be performance issue with database?
Comment