https simple check sometimes reports service down

[tvtue]

Hi all,

the https simple check sometimes reports "service down" but when I log in to the server I see that apache is running and port 443 is open. When I look in apache's access log, I can see some checks are missing in the five minute intervall.

Is the zabbix server to slow? How can I debug this?

Regards,
tvtue

[Heilig]

Please, check http poller processes utilization by using App Zabbix Server Template.

[tvtue]

Hello Heilig,

thank you for your reply.

The average value over one month in the graph of busy unreachable poller processes lies at 2.76% and it was zero in one of the events where https was reported to be down.

Regards,
tvtue

[Heilig]

Is poller processes utilization also low?
Is not there network delays when receiving results?
Maybe you should make the trigger less sensitive (by modifying its expression)?

[tvtue]

Thank you very much for your hints!

The "busy http poller processes" graph over one month period shows an average of 0.08 % and a peak of 20.91%. And exactly on one event where it happened, the value was also at 0.2%.

I also have a ICMP ping item setup to that server. It show no slow down in the ping response time at that certain time.

Ok, I could loosen the trigger expression by saying the last 3 must be 0 or so. But that would be a last way out.

tvtue