Gaps in agent data graphs only only on some hosts

[r0k]

Dear All,

today I was looking through some agent data graphs and noticed that on some hosts (SLES 15) there are huge gaps in the graphs. At the same time on (CentOS 6/7) there are no gaps and everything looks perfectly fine.

The monitored nodes are bare-metal and virtual ones. They run the same version of the zabbix agent zabbix-agent-4.4.10-1 and they are monitored by the same proxy. I enabled debug logging on the clients and the proxy but found noting. I used zabbix_get and pulled e.g. agent.ping every second and I always get an imidiate reply.

The zabbix server and proxy version is 5.0.11. Even the config files on the clients are 100% identical. We use puppet to deploy and manage zabbix.

I know that the agent version should match the server so I updated the agent on the SLES 15 VM where I have gaps, but still no change. I have the feeling that only SLES 15 machines are affected. But I could be wrong, at least I have not found one CentOS machine yet that shows the same behaviour.

Any help is highly appreciated.

[r0k]

Dear All,

problem seems to be solved. There was an issue with the VMware Cache size being to low and due to this the Zabbix Server would crash every now and then and not collect data.

egrep "__mem_malloc|mem_realloc" /var/log/zabbix/zabbix_server.log | head

Best Regards,
Oliver