So I have this running in my test environment with about 100 hosts, all doing about 30 'zabbix agent' checks. I decided to change some (network stats, processor stats) to 'zabbix agent (active)' to see if it would help out load wise.
Instead what happened is that the queue filled up with requests stating that the next time to update the check was 12/31/1969 even after restarting the server. Problem with that is that date is before epoch so I can't ever see it complete. The process queue load shut up from 4-5 (on a 4 cpu machine) to 24-25... not pretty.
Instead what happened is that the queue filled up with requests stating that the next time to update the check was 12/31/1969 even after restarting the server. Problem with that is that date is before epoch so I can't ever see it complete. The process queue load shut up from 4-5 (on a 4 cpu machine) to 24-25... not pretty.