Hello Zabbix team!
I noticed in my server log a ton of error messages back to back. These messages had flooded my log file:
13000:20081008:103543 Format error or unsupported operator. Exp: [closer]
13000:20081008:103543 Expression [({12593}=1)&({12592}>0)] cannot be evaluated [Format error or unsupported operator. Exp: [closer]]
13000:20081008:103613 Format error or unsupported operator. Exp: [closer]
13000:20081008:103613 Expression [({12593}=1)&({12592}>0)] cannot be evaluated [Format error or unsupported operator. Exp: [closer]]
Those messages didn't help identify the trigger, the item, nor the host that this problem existed on. I studied the log and realized that the messages were every 30 seconds and had 2 parameters in the trigger field. Since I had less then a dozen systems on this particular zabbix server, I eventually narrowed it down to the item that was having this problem and we realized that the item was sending a value that we had not expected.
I also realize that this trigger was listed as 'Unknown' while it was in this state, but unfortunately it was on a dev box that we test a lot of software and alerts on...there are always a few triggers in the 'Unknown' state.
I realize that had I been diligent and continuously flushed out all the old test triggers on this box, the 'Unknown' would have stood out. It was also a problem on my end that we didn't account for all the values being sent to Zabbix. I am sure that turning up the log verbosity would have also yielded a result. However, it would have really been nice to have some identifier in the standard log file. Hostname, trigger, item, anything would have been nice to help better identify the problem trigger.
So my request is that an identifier of some sort be used for problem areas in the log file.
Thanks!!
I noticed in my server log a ton of error messages back to back. These messages had flooded my log file:
13000:20081008:103543 Format error or unsupported operator. Exp: [closer]
13000:20081008:103543 Expression [({12593}=1)&({12592}>0)] cannot be evaluated [Format error or unsupported operator. Exp: [closer]]
13000:20081008:103613 Format error or unsupported operator. Exp: [closer]
13000:20081008:103613 Expression [({12593}=1)&({12592}>0)] cannot be evaluated [Format error or unsupported operator. Exp: [closer]]
Those messages didn't help identify the trigger, the item, nor the host that this problem existed on. I studied the log and realized that the messages were every 30 seconds and had 2 parameters in the trigger field. Since I had less then a dozen systems on this particular zabbix server, I eventually narrowed it down to the item that was having this problem and we realized that the item was sending a value that we had not expected.
I also realize that this trigger was listed as 'Unknown' while it was in this state, but unfortunately it was on a dev box that we test a lot of software and alerts on...there are always a few triggers in the 'Unknown' state.
I realize that had I been diligent and continuously flushed out all the old test triggers on this box, the 'Unknown' would have stood out. It was also a problem on my end that we didn't account for all the values being sent to Zabbix. I am sure that turning up the log verbosity would have also yielded a result. However, it would have really been nice to have some identifier in the standard log file. Hostname, trigger, item, anything would have been nice to help better identify the problem trigger.
So my request is that an identifier of some sort be used for problem areas in the log file.
Thanks!!
Comment