JMX Graphs with gaps

[ftclausen]

Hello All,

First, some information

Server Version: 2.2.1
Server Platform: RHEL 6
Agent Version: 2.2.1
Agent Platform: RHEL 5

I am having a peculiar problem with a host under monitoring (and only that particular host) whereby JMX graphs have gaps. See graph below for an illustration



The other, non JMX graphs are OK. I have been running jmxterm in a loop to see if there are any errors during collection and jmxterm does not seem to miss a beat. Other hosts under monitoring using JMX do not exhibit this problem.

The Zabbix server health graphs imply things are OK :

* Zabbix server performance - nothing in the queue
* Zabbix data gathering process - around 5% busy
* Cache hits look good
* Cache usage low
* PS output shows the java pollers are mostly idle
* No related errors/messages in gateway logs

I know it is a long shot since it is probably something specific to my environment but I thought I'd ask anyway - in case someone has experienced this before.

Many thanks,

Fred.

[EnigmA-X]

I had a similar issue once, some time ago. In the end it turned out that I was reaching the ulimits set on the system.

It's a long shot, but you might want to check your ulimits for the zabbix user.

Check /etc/security/limits.conf and /etc/sysctl.conf

[ftclausen]

I had a similar issue once, some time ago. In the end it turned out that I was reaching the ulimits set on the system.

It's a long shot, but you might want to check your ulimits for the zabbix user.

Check /etc/security/limits.conf and /etc/sysctl.conf

Thanks! I will check that out.