-
Incident report
-
Resolution: Commercial support required
-
Trivial
-
None
-
None
-
None
-
Red hat 7.9 - Zabbix Server 5.0.6
Hi everyone, we're running into a problem in zabbix that we haven't been able to resolve yet. We will try to explain it in detail:
We have a zabbix that has a snmp listener and receives traps, these same ones go through /etc/zabbix/zabbix_trap_receiver.pl
and stores its output in /var/log/zabbix_traps.tmp. A host was created to be able to read
these traps and a corresponding trigger. The trigger has a TAG to analyze the alarm ID
and I established if it is a problem or a recovery. At the time that a new state is provoked,
either a problem or an OK, it triggers the execution of a script to which they are passed
the received alarm data is sent via SNMP to a TeMIP server (zabbix works as "forwarder")
see the capture "item"
see the capture "trigger"
see the capture "trigger tag"
Carrying out different tests, we see that having more than one problem alarm simultaneously (for example 3), they arrive correctly
to snmptrap (log in /var/log/zabbix_traps.tmp) and then the script is executed with the data of the corresponding alerts. Until here the
circuit works correctly. The problem lies when the OKs are established. When solving those 3 "problems" of the detailed example,
We verify that the traps arrive correctly to the snmptrap (/var/log/zabbix_traps.tmp), but when the return of the zabbix is ​​established
so that the script with the resolutions is then executed and sent to the Temip server, we find the following:
Of the 3 resolved events, the first is displayed with the status "problem" instead of OK, the second event is omitted, and the third has the satus OK, as appropriate,
but it is repeated 2 times.
see the capture "problem"
see the capture "ok"
See the file "problem traps zabbix.tmp"
Here is how the traps arrive with status problem in zabbix.tmp. This is correct
See the file "forward problem traps"
after executing the action of the script to forward the traps to the Temip server
See the file "OK traps zabbix.tmp"
This is how traps with status OK arrive in zabbix.tmp. This is correct
See the file "forward ok traps"
Here you can see how the first event continues with the status "problem" when in reality it has already been resolved. You can also see that it ignores the second event and the third repeats it 2 times.
Can someone guide us to what could be the cause of this problem? The scenario is repeated having from 2 simultaneous resolutions onwards.
Thank you!