I've setup a python script that uses the Zabbix API running once a day in cron to iterate over each zabbix host and for each "Host Unreachable" trigger set a dependency on that hosts Zabbix "Proxy Timeout Exceeded" trigger, which is a trigger on the Zabbix server itself. This works as expected until recovery.
1) Disconnect the environment monitored by proxy (proxy and all hosts behind it are now unreachable from the Zabbix server).
2) Only the proxy timeout trigger escalates (dependencies working as intended)
3) Wait a period of time, doesn't matter how long (tested 5-30 minute breaks)
4) Reconnect the environment monitored by proxy
5) At the time of event #4 the proxy timeout trigger sends it's OK escalation, but then all "host unreachable" triggers behind that proxy perform their critical escalation actions as well as sending their okays.
#5 is not working as expected. What I would expect, and hopefully you can help, is for only the proxy timeout to send it's okay. The host unreachables all trigger critical right when the proxy re-connects.
Can someone please tell me what I may be missing here or if this is a bug in the way dependencies work?
1) Disconnect the environment monitored by proxy (proxy and all hosts behind it are now unreachable from the Zabbix server).
2) Only the proxy timeout trigger escalates (dependencies working as intended)
3) Wait a period of time, doesn't matter how long (tested 5-30 minute breaks)
4) Reconnect the environment monitored by proxy
5) At the time of event #4 the proxy timeout trigger sends it's OK escalation, but then all "host unreachable" triggers behind that proxy perform their critical escalation actions as well as sending their okays.
#5 is not working as expected. What I would expect, and hopefully you can help, is for only the proxy timeout to send it's okay. The host unreachables all trigger critical right when the proxy re-connects.
Can someone please tell me what I may be missing here or if this is a bug in the way dependencies work?