Hello
We are facing an issue with defining triggers to avoid false recovery of triggers.
The situation is as follows
We are monitoring the RX power on a number of discovered interfaces using SNMP, and want an alert when the value changes. This is easy.
The recovery should however only happen when the value goes back to something that can be consider normal. Also easy.
However, due to the nature of the equipment we are monitoring there is no default range of RX power values we can set as condition for what should be normal since that is unique for each interface. This is what we get stuck on.
Currently when the value changes we receive a problem alert, then when the next check happens from zabbix and the value hasn't changed again an ok is sent. This is not the behaviour we wish to achieve
We instead wish to have it send a problem alert on the change in RX power, then remain in problem status until the value goes back to what is normal for that specific interface.
The first thought I had was to have the trigger compare to what the value was X hours ago along with a second condition to eliminate small value changes caused by natural fluctuations, but it does not seem like an intuitive and robust way of solving the problem
Any ideas or suggestions for how such a scenario could be solved?
We are facing an issue with defining triggers to avoid false recovery of triggers.
The situation is as follows
We are monitoring the RX power on a number of discovered interfaces using SNMP, and want an alert when the value changes. This is easy.
The recovery should however only happen when the value goes back to something that can be consider normal. Also easy.
However, due to the nature of the equipment we are monitoring there is no default range of RX power values we can set as condition for what should be normal since that is unique for each interface. This is what we get stuck on.
Currently when the value changes we receive a problem alert, then when the next check happens from zabbix and the value hasn't changed again an ok is sent. This is not the behaviour we wish to achieve
We instead wish to have it send a problem alert on the change in RX power, then remain in problem status until the value goes back to what is normal for that specific interface.
The first thought I had was to have the trigger compare to what the value was X hours ago along with a second condition to eliminate small value changes caused by natural fluctuations, but it does not seem like an intuitive and robust way of solving the problem
Any ideas or suggestions for how such a scenario could be solved?