Resolve how to handle case when the handler crash

[MalloZup]

it can be solved in many ways.

An option would be to have somehow kind of redundancy on the daemon itself which will respawn and monitor a sub-daemon.
In case it fails it will be respawned.

Since the alert-handler can be fondamental component, I do think that somehow a similar solution where only this binary is involved for its own redundancy/resilience is the best solution