Immediate actions upon problem detection
At any time each monitor can be in one of three states: Ok, Warning and Error. A monitor can change its state after each probe. You can specify the interval between probes and conditions required to switch between states. For example, you can specify normal ping time limit. If a ping monitor returns longer time, it switches to the “warning” state. You can also specify a timeout value that would put it to the “error” state.
The Warning state is useful to identify various bottleneck problems on early stages. If you notice that some monitor switches to that state from time to time and returns back to “Ok”, it may suggest a required hardware upgrade or signal some inefficiency in the system.
When the state of any monitor changes, the program can do any of the following actions.
For each monitor you can specify any set of actions to be performed on any change of its state. It is also possible to launch any of these actions after several signals from the same monitor. This is useful to filter false alerts.
- Send immediate notification by email, SMS or an IM service.
- Launch another application that can resolve the problem. For example, it can restart a server or some service on it, enable reserve service.
- Run or stop another monitor or group of monitors. This is useful to create dependencies between different monitors and groups.