In the course of investigating both a sudden increase in memory usage and a "LLD Poller over 75% Utilisation" alert that popped up on my Zabbix 6.0 server (which I don't necessarily think are directly related, judging by timeframe), I noticed that the LLD processes seem to run unusually hot on CPU, they're pretty much using an entire core each near constantly, and this persists even after adding a couple more pollers to clear the 75% utilisation alert. Most poller types don't run anywhere near this hot on resource, so I'm not sure fi this is normal or if there's an issue or misconfiguration somewhere I need to try and troubleshoot. There doesn't seem to be a lot in the Zabbix server logs that references LLD, even with debug levels turned up that might point to a problem.
Nobody with access to the server has come forward when asked about adding any new templates or batches of hosts etc.
Is anybody able to advise whether this is normal behavior or not, and how I might investigate further if this is not expected?
Many thanks.
Code:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1694793 zabbix 20 0 4737756 119052 41384 R 99.7 0.0 1978:09 zabbix_server 1694790 zabbix 20 0 4752620 132400 40116 R 99.3 0.1 7821:00 zabbix_server 1694792 zabbix 20 0 4755760 139364 43668 R 81.8 0.1 5534:13 zabbix_server
Is anybody able to advise whether this is normal behavior or not, and how I might investigate further if this is not expected?
Many thanks.
Comment